2026 Collaboration Cafe Notes Archive#

2026-09-01#

Check-in 💁#

Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Min (he/him) / @minrk / UC Berkeley / California

  • Kirstie (she/her) / @KirstieJane / UC Berkeley / California

  • Ryan Lovett (he/him) / @ryanlovett / UC Berkeley / California

  • April Johnson (she/they) / @aprilmj / 2i2c / Virginia

  • Serena Bonaretti / @sbonaretti / Switzerland

  • Kelly Rowland / @kellyrowland / NERSC/ Calif.

  • Harold Campbell / @haroldcampbell / 2i2c / Jamaica

  • Jenny Wong (she/her) / @jnywong /2i2c/ Yorkshirrrrrre

  • Stéfan van der Walt / @stefanv

  • Angus Hollands (he/him) / @agoose77 / 2i2c / Rugby, UK

  • Freek Pols / freekpols / Delft University of Technology

  • Faith Uniter Nyakerario/ @afterthesegfault / NAILUG / Nairobi, Kenya

Ice breaker 🧊⛏️#

What was your first programming language or first “technical obsession”?

  • Min: TI-82 (calculator) BASIC

  • Kirstie: I learned a sed one liner (https://www.unixguide.net/unix/sedoneliner.shtml) at a summer school at the end of my first year of my PhD and it (combined with a bash for loop) transformed my LIFE!

  • Ryan: BASIC on a Commodore, PET I think

  • April: COBOL (for work) and PASCAL for school I think?

  • HMC: GW-Basic and first technical obsession was writing terminate-and-stay-resident (TSR) programs in Pascal

  • Jen: VBA Excel lol

  • Stéfan: discovered a Turbo Pascal manual and typed in all the graphics examples

  • Angus: PHP 🥹 Just wanted a guest book on my website

  • VBS

  • First programming language, Golang, Technical obsession the code that made Architectural software like Archicad run.

Celebrations and shout-outs 🎉#

Breakout room and agenda item suggestions#

Notes for Main room:#

  • We are discussing planning for Hub Dash

  • Dates coincide with The Turing Way Book Dash

  • Send out a survey before the event and disseminate results at the event

  • Last time, there was a call for proposals for what to do at the event

    • e.g. a list of suggested topics and participants ranked them in the order they preferred

  • Serena will propose a draft timeline of things for us to work on to prepare for the Hub Dash

  • Serena has also started a draft of the survey

    • We spent a few minutes individually looking at it and making small edits

    • Then we started looking at it together section by section

    • Todo: privacy policy for GDPR/Data compliance

      • ask respondent if user is happy to share detailed answers attributable to thempublicly/publish answers in aggregate

    • General tip: give multiple-choice answers with an Other option to make analysis easier

    • Ryan: are there any questions from the Organizational Mycology report that could be relevant to ask to a broader audience in this survey?

2026-08-18#

Check-in 💁#

Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Freek Pols / freekpols / TU Delft / Netherlands

  • Serena Bonaretti / @sbonaretti / Switzerland

  • Jérôme Lecoq / jeromelecoq / Allen Institute / Seattle, WA

  • Chris Holdgraf / @choldgraf / 2i2c / Oakland, CA

  • Silas Santini / @pancakereport / UC Berkeley

  • Kirstie Whitaker (she/her) / @KirstieJane / UC Berkeley / Berkeley, CA

  • Harold Campbell /@haroldcampbell / 2i2c / Jamaica

  • Arielle Bennett / @Arielle-Bennett / The Alan Turing Institute / Boston, MA

  • Kelly Rowland / @kellyrowland / NERSC / Oakland, CA

  • Faith Uniter / @afterthesegfault / NAILUG / Nairobi, Kenya

  • April Johnson / @aprilmj / 2i2c / Richmond, VA

Ice breaker 🧊⛏️#

What is the last thing you learned unexpectedly?

  • Kelly - I learned that in older nautical times they said “larboard” and “starboard” for left and right on the ship, respectively. Larboard was changed to “port” at some point because it is difficult to hear the difference in, e.g., a squall.

  • To connect large Ikea cabinets

  • To wash the car in Switzerland!

  • How thunderstorms have warm air that shoots upward up to 60 mph surrounded by colder air. I knew warm raises, but I didn’t know it would go so fast

  • The secret backstory of how a locally iconic bar was saved

  • We have an entire paper in the making on unexpected responses in the brain. Look it up: https://allenneuraldynamics.github.io/openscope_p3_data_release_paper/

  • The psychology behind the mean nicknames we give our loved ones.

Celebrations and shout-outs 🎉#

Breakout room and agenda item suggestions#

  • Serena: JupyterHub: Proposal for the 2nd HubDash: Creating a collection of use cases of JupyterHub deployments (jupyterhub/team-compass#926)

  • Demo from Jerome: DeMystify

Notes for Main room:#

2026-08-04#

Check-in 💁#

Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Kirstie Whitaker (she/her) / @KirstieJane / UC Berkeley / Berkeley, CA (currently at Lawrence Berkeley National Lab looking out over campus and the bay!)

  • Faith Uniter Nyakerario/@afterthesegfault/ NAILUG/ Nairobi,Kenya

  • Arielle Bennett / @Arielle-Bennett / The Alan Turing Institute / Boston, MA

  • Min RK / @minrk / UC Berkeley / California

  • Silas Santini / @pancakereport / UC Berkeley

  • Serena Bonaretti / @sbonaretti / Switzerland

  • Kelly Rowland / @kellyrowland / NERSC / SF Bay Area

  • Harold Campbell/ @haroldcampbell / 2i2c/ Jamaica

  • Yijun Ge /@yijunge-ucb / UC Berkeley

  • Brigitta Sipőcz / @bsipocz / Caltecht

Ice breaker 🧊⛏️#

What project (technical or non-technical) are you currently excited about?

  • Kirstie: Lightcone Research: https://lightconeresearch.org

    • A new open source project here at BIDS (and at CNRS in Paris, France) working to ensure agentic workflows are auditable and trustworthy

    • We ran their Developer Summit at BIDS’s AI Futures Lab last week and it went SO WELL!

      • Loved meeting Deb from NERSC and talking with her at LENGTH about JupyterHub!

  • Faith:Building butterboxes and interacting with raspberry Pis also my talk was accepted at ATO

  • Min: Running ethernet under my house to improve signal in my home office, which is currently far from my router. We’ll see what the rats and frogs think about it. I am also doing this, but by “doing it” I’m asking someone else to because my walls are solid af

  • Silas: working my way up to longer and longer open water swims

  • Re-launching an opensource project I started years ago

  • Kelly: started reading Moby Dick, big classic lit summer

  • Arielle: working on my most detailed paint-by-numbers yet (it’s a repro of a scientific illustration!): https://canvasbynumbers.com/products/actinae-ernst-haeckel-paint-by-numbers-kit

    • Sick

  • Angus: I’m upgrading my home network on top of OpenWRT

Celebrations and shout-outs 🎉#

  • Serena for engaging with / proposing JHub/JBook meeting changes!

Breakout room and agenda item suggestions#

Notes for Main room: JupyterHub Contributing Guidelines#

2026-07-21#

Check-in 💁#

Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Min (he/him) / @minrk / UC Berkeley / California

  • Ryan Lovett (he/him) / @ryanlovett / UC Berkeley

  • Yijun Ge (she/her) / @yijunge-ucb / UC Berkeley

  • Freek Pols / @freekpols/ Delft University of Technology, Netherlands

  • April Johnson (she/they) / @aprilmj / 2i2c / Virginia

  • Jenny Wong (she/her) / @jnywong / 2i2c / Yorkshirrrrre, UK

  • Mike Krassowski / @krassowski / OpenTeams / Usually UK, today Poland

  • Kelly Rowland / @kellyrowland / NERSC / Calif.

  • Faith Uniterh/@afterthesegfault /NAILUG/ Nairobi Kenya

  • Chris Holdgraf / @choldgraf / 2i2c / Emerrrrrrrrrrrrrrrryville, USA

  • Serena Bonaretti / @sbonaretti/ Switzerland

Ice breaker 🧊⛏️#

What is one tool, app, or workflow you discovered recently that you really like?

  • How to use AI with notebooks, MyST docs

  • I started using direnv in combo with tmux recently and I like it!

Celebrations and shout-outs 🎉#

  • Dawa Ometto for landing unix socket support across jupyterhub

  • Kubespawner release. Thank you @GeorgianaElena and @manics!

Breakout room and agenda item suggestions (& notes)#

2026-07-07#

Check-in 💁#

Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Min (he/him) / @minrk / UC Berkeley / California

  • Arielle (she/her) / @Arielle-Bennett / The Alan Turing Institute / Boston, MA

  • Kirstie (she/her) / @KirstieJane / UC Berkeley / California

  • Michael / @mforbes / Washington State University / Washington

  • Justin Lubin (he/him) / justinlubin / UC Berkeley / California

  • Yuvi / @yuvipanda / 2i2c / California

  • Kelly Rowland / @kellyrowland / NERSC / Calif.

  • Brigitta Sipőcz / @bsipocz / Caltech / Seattle, WA

  • Matt Fisher / @mfisher87 / Schmidt DSE / Longmont, CO

  • Stéfan van der Walt / @stefanv / UC Berkeley-BIDS-SP

Ice breaker 🧊⛏️#

What song, podcast, or YouTube channel have you been enjoying lately?

  • [min] Worlds Beyond Number: The Wizard, the Witch, and the Wild One +100 this is an excellent story!

  • [Arielle] It’s been all football all the time in this house, so please enjoy a classic English football song, Fat Les - Vindaloo https://youtu.be/va6nPu-1auE?si=NM3Vt-t2vrdhPZbm + 1

  • [Kirstie] Karen Puzzles - speed puzzling!! https://www.youtube.com/@KarenPuzzles

  • [Eric] Podcast: the worst idea of all time – two comedians watch the same movie once a week for a year. Its horrific and beautiful

  • [Brigitta] - nthe podcast I’m currently into are the Empire podcast (currently on season 5), The Avalanche hour

  • [Angus] unironically the Bangladesh song on TikTok

  • [Yuvi] Lots of Ed Zitron :) Also Flight of the Conchords

  • Love Ed’s writing on AI!

  • [Michael] People I (mostly) admire (Freakonomics) https://freakonomics.com/series/people-i-mostly-admire/

  • [Matt] Podcast: Change, Technically!!!! https://www.changetechnically.fyi/

Celebrations and shout-outs 🎉#

Notes for Breakout room 1: JupyterHub overview#

  • Arielle shared the notes from the Users of JupyterHub workshop form JupyterCon: https://annuel.framapad.org/p/jupyterhub-users-jc-nov-2025-ahog

  • Justin & Eric will be talking with + observing computational biologists this summer

  • Eric: Some goals for the meeting:

    • (1) Get a shared understanding of the “vocabulary” of JupyterHub

    • (2) Come up with set of questions of interest that Eric/Justin could help answer

    • Identify what’s stopping people from using JupyterHub

  • Kelly: Curious why interested in JHub for computational biologists? I think of JH as a vehicle rather than the end-user product that the user wants to interact with directly.

    • Eric: Also a question I have! Infrastructure concerns are often dealbreakers for users. Identifying those sticking points that can be difficult to see from the outside. I’m interested in JupyterLab as a programming interface. How are folks using these interfaces? What are the opportunities for Notebooks in general?

  • Yuvi:

    • Looking at JH from lens of accidental complexity. Programming is a tool, and requires a lot of arcane knowledge to engage with (ssh, vim, other terminal things, job queue). “I just wanted to read the DB”. This all seems normal to people who are steeped in it, but alien to those who aren’t. JH is infrastructure that should be invisible and help solve problems that add an accidental complexity “tax” to the real work people want to do. “We are felt by our absences” (i.e. things you don’t need when you use JH. Like to learn to SSH. Or as a workshop organizer, to deal with unique OS/software/hardware issues from participants)

    • Talk @ BIDS: Democratizing everything with wikipedia principles <TODO: Link>

  • Min:

    • Goals of JH are about, when you have access to computing resources, how do you make the right choices for your end users. What do you want them to be able to do? And what do you want to restrict them from doing? (convenience & security reasons)

      • Installation

      • Homogenous environments

      • Environment customization (tension between ability to break environment and needing to do admin requests for every package they need)

    • JH users are infrastructure engineers. JupyterLab users are the folks doing analysis in a Notebook or whatever.

    • Design questions center around what do we want to enable users to do, and how do you make that as smooth as possible, and what do you want users to understand or not understand

  • Matt: design tension between smoothing out problems and teaching portable skills?

    • Min: Try to do things that are unique or special as little as possible

    • Min: What do you do for users? E.g. shipping an environment may mean users don’t think of their dependencies because they’re already there.

    • Min: We think of this as out of scope – we allow infrastructure people to make those choices! They can balance that tension. E.g. complete control: BYO image; or tightly controlled: one pre-made image.

    • Min: We demonstrate best practices (we only have a little bit of this)

    • Min: Avoiding opinionated one-way-to-do-thing decisions!

    • Yuvi: Portable skills is not just about moving from hosted -> local, but also doing the same thing between different hubs. The answer about portability should be a social answer, not technical. E.g. you can’t get GPU on this hub but not another hub, ask the admin why!

  • Eric: Rephrasing, formulating design questions for JupyterHub:

    • How do we support the wide diversity of ways programmers are working while still giving the JupyterHub maintainers flexibility to add guardrails

      • Matt: Terminology check – when we say “maintainers” do we mean folks who manage the JupyterHub codebase? And when we say “admins” we mean folks who are managing deployments of JupyterHub?

  • Yuvi:

    • Most useful thing to know: What is the accidental complexity that computational biologists still have to deal with? E.g. anything that they say “I’m doing this, I wish I wasn’t doing this…”

      • Being able to notice the specific parts that we can fix

  • Eric: Anxious about technical vocabulary. But it sounds like that’s not what’s needed. What’s needed is knowing about sticking points and then iterating together to solve them!

  • Justin: What sorts of accidental complexity is in scope? Is it anything that gets in the way of science workflows, or specifically infra issues?

    • Min: Would put the limit at domain-specific stuff. E.g. complexity of configuring a specific biology analysis package would be out of scope. We care about a certain level of generic-ness. Can be case-by-case.

      • Matt & Yuvi: Given case-by-case-ness of whether stuff is in scope, be communicative about your problems – if they don’t fit JupyterHub, we (you, Min, Kelly, Yuvi, Arielle, Angus, etc.) can probably figure out together which part of the Jupyter community might be interested in solving it.

      • Yuvi: “Discovering the generic-ness together!” <3

  • Some specific questions:

    • Yuvi:

      • Where are the R people at? What interfaces/infrastructure are they using? How are they doing package management? What is their workflow / their pain points?

  • Min:

    • “What are you doing that you want to not be doing?” (Highlighting what Yuvi said earlier)

      • This is really what a lot of the infrastructure projects are about - taking something that is annoying, frustrating, and common and enabling someone to solve them so they can do the actual work they need to do

        • The activities/responsibilities that people wish were not theirs

    • Yuvi: An example of this: Data 8 was struggling with getting the homework notebooks to each student, make changes if needed, and send them out. They had their own bespoke solution for this that didn’t work perfectly - the generic version of this problem though is “we want to be able to clone a GitHub repository without knowing about git” (even though that’s not how they would have described it)

      • People always have lots of problems… to be able to see what is critical and what is not is very valuable. When did people have to give up / what is the point at which you had to give up? Not just what was hard, but the failure cases (e.g. SSH used to be a really big one before).

        • Usually the point at which the person gives up is not the point at which you introduce the fix - the fix is much earlier

  • Eric: What should our writeup/deliverable/output look like? (Justin: i.e. what would be most helpful for you?)

    • Synthesizing insights/themes from video recordings of actual users?

    • “Vignettes”

    • Matt (not on the JH maintainer team, just my thoughts :) ):

      • Something I’ve found useful is something like an “experience report issue” type in GitHub - ie a set of the exact steps a user took that demonstrates pain points

    • Yuvi:

      • It would be really helpful to have timepoints of the raw video [may have privacy considerations to work around though].

        • Eric: Could look into putting Yuvi on the IRB

          • Yuvi: Also, can ask Eric follow-up questions about anything in the vignettes

  • Yuvi: Highly recommend working with the JupyterLab folks too!

    • Min: The JupyterLab meeting is tomorrow (Wednesday), would probably be excited to talk with you!

Notes for Breakout room 2: Jupyter Book Renaissance Grant Proposal#

Original LOI submitted: jupyter-book/renaissance2026-myst

Please note that you’ll have the ability to make adjustments to your proposal or to expand your proposal’s scope in the final full application. However, we request that the overall focus of the application remain aligned with the original LOI. https://os4science.org/funding_opportunity/os4ls/ https://os4science.org/wp-content/uploads/Full-Application-Instructions_-OS4LS.pdf

  • Angus +1 on JupyterHub & MyST

    • Different to optimise Jupyter Book / MyST for the browser vs a nodejs runtime

    • Allows us to make MyST more modular as a “side effect”

    • Greater API surface - allows us to have more plugins too

  • Angus +1 on integrating MyST better into JupyterLab

  • Command line tools useful for AI skills

  • Editor integration? VSCode, LSP, highlighting,

  • Integration with ingest/tooling on ipynb files.

  • Notebook flavors

  • Markdown flavors

  • Can we position this grant as a fully end to end scientific communication

    • JupyterHub + JupyterLab + JupyterAI + Jupyter Book for publication

    • Needs MyST to work in JupyterHub and JupyterLab effectively

  • What’s the goal for improving our plugin architecture

    • Community created

    • Allow others to interoperate

    • Wouldn’t make sense for MyST to be curated to meet all life science needs - but we need to improve the architecture so that others can interoperate with MyST

      • can we find a couple of examples here

  • GOALS

      1. Refactor MyST so that it works effectively in JupyterLab

      • Full loop: AI + lab + myst (computational narrative notebook -> publication)

      • Enable smaller “parts” of myst that do specific things – tool calling

      1. Refactor MyST so that it works effectively in JupyterHub

      • Not that much to do specific to Jupyter Hub

      • Adjust this to a Publishing workflow

      1. Enable community participation by surfacing more options to connect plugins

      • What are we enabling that is not currently possible?

      1. Enable community participation by surfacing more options to integrate widgets

      • What are we enabling that is not currently possible?

      1. Run events? 2 x developer summits? w user input?

      • Highlight Chiara’s recent workshop!!

2026-06-16#

Check-in 💁#

Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Min (he/him) / @minrk / UC Berkeley / California

  • Him/he Emmanuel Odenyire Anyira/WorldQuant University/US

  • Dawa (he/him) / @dometto / Utrecht University / Netherlands

  • Kirstie (she/her) / @KirstieJane / UC Berkeley / California

  • Luciano Resende / @lresende / Apple / California

  • Kelly Rowland (she/they) / @kellyrowland / NERSC / Calif.

  • Silas Santini / @pancakereport / UC Berkeley

  • Ryan Lovett / @ryanlovett / UC Berkeley

  • April Johnson / @aprilmj / 2i2c / Virginia

  • Faith Uniter/ @afterthesegfault/NAILUG,FDroid/ Nairobi,Kenya

Ice breaker 🧊⛏️#

Mountains, sea, forest, or city?

  • Min: Mountains > Forest > City > Sea

  • Kirstie: Mountains > Forest > City > Sea

  • April: Sea > City > Forest > Mountains

  • Ryan: Forest > Sea > City > Mountains

  • Silas: City > Mountains > Sea > Forest

  • Dawa: Sea > Mountains > City > Forest

  • Faith: Sea> City > Forest > Mountain

Celebrations and shout-outs 🎉#

Breakout room and agenda item suggestions#

  • PR Review: Dawa Ometto, unix socket connection to CHP

  • jupyterhub idle culler release?

    • Min suggested looking at the list of changes (via github-activity) in order to help determine whether the release is a major, minor, or micro version bump. That tip could/should be added to the RELEASE.md.

    • It is important for maintainers to apply labels to PRs. This could also be done before running github-activity, and also after while iterating, in order to improve its output

    • Min sometimes adds to the changelog addition PR description a note about when he’ll merge it

    • Issue around main tracking might need to be documented

      • git branch –set-upstream-to upstream/main main

  • Review of security advisory issues

Notes for Breakout room 1: Main room - PR review#

  • jupyterhub/jupyterhub#5397

    • Submitted by Dawa Ometto, reviewed by Min RK

    • Dawa’s notes on the context for this PR: our researchers work on virtual machines that are shared between multiple users. These machines run JupyterHub behind Nginx. Nginx redirects to a separate single-sign on (OAuth) server; if authentication is successful, a REMOTE_USER header is set with the username provided by the OAuth server and passed on by Nginx to JupyterHub (i.e. this pattern: cwaldbieser/jhub_remote_user_authenticator). This works fine for HTTP requests that originate from outside the VM. However, if CHP is listening on a TCP socket, then users who are logged in locally on the VM can curl http://localhost:8000 and add an arbitrary REMOTE_USER header; JupyterHub has no way of knowing that the header does not originate from the single sign-on server. Hence the importance of letting CHP listen on a unix socket instead. The current PR to JupyterHub is to make sure that CHP, when managed (started/stopped) by JupyterHub, can actually utilize CHP’s ability to listen on a unix socket. Sidenote 1: this isonly an issue given that we’re using shared VMs and using HTTP headers for authentication – there would be no issue e.g. if the Hub was spawning separate containers. But the cloud platform (custom made for Dutch/European universities) we’re using at the moment is VM only (no k8s), and we can’t easily switch to a different form of authentication. So though our usecase is a bit specific, I hope others who are using JupyterHub on shared machines can also benefit from some more security.

  • Kelly’s first release!! jupyterhub-idle-culler 2.0.0

2026-06-02#

Check-in 💁#

  • Min (he/him) / @minrk / UC Berkeley / California

  • Yann (he/him) / @Yann-P / Independant / France

  • Kirstie Whitaker (she/her) / @KirstieJane / Berkeley Institute for Data Science / California, USA

  • Serena (she/her) / @sbonaretti / Switzerland

  • Faith Uniter (she/her) / @afterthesegfault / Nairobi, Kenya

  • Harold Campbell / @2i2c/ Jamaica

  • Brigitta Sipőcz (she/her) / @bsipocz / Caltech / Seattle, USA

  • Silas Santini / @pancakereport / UC Berkeley

  • Angus Hollands / @agoose77 / 2i2c / Rugby, UK

  • Ryan Lovett / @ryanlovett / UC Berkeley

  • Simon Li / @manics / Uni Dundee

  • Stéfan van der Walt / @stefanv / BIDS+UCB+SP

Ice breaker 🧊⛏️#

What book (or other thing that you can read with your eyes or ears) are you enjoying at the moment?

  • Min: latest Murderbot

  • Unreasonable Hospitality: https://www.unreasonablehospitality.com/books

  • Deep work by Carl Newport (Harold: This was a really good book. I loved this one!)

  • depression hates a moving target – I forgot my kindle for a camping trip, a friend had this as a second book and I stuck with it

  • Doppelganger by Naomi Klein

  • Yann : Legend by David Gemmell

  • Angus: blog on accessibility in Wayland https://nocoffei.com/?p=451

  • Serena: an Italian book! by F. Volo

  • Ryan: A Parade of Horribles

Celebrations and shout-outs 🎉#

  • Delighted to have Serena on the team! So great to work with you!

Breakout room and agenda item suggestions#

  • JupyterHub PR Review (with Min this time)

  • Roadmap item done-ish: jupyterhub/roadmap#8

    • just needs blog post…

  • Jupyter Book:

    • Talking about a Python Jinja theme?

    • I want to know about developer docs and things like “definitionList”

Notes for Breakout room 1: Jupyter Hub#

Notes for Breakout room 2: Jupyter Book#

  • A lot of PR review

    • Bugfixes

    • Theme release via GH Releases

    • Moving to React 19 and bun!

  • Discussion about purpose of myst-theme and mystmd theme docs

  • Proposal to build out MyST theme that uses Jinja templates and Python (for maximal developer familarity)

2026-05-19#

Check-in 💁#

  • Kirstie Whitaker / @KirstieJane / UC Berkeley / Berkeley, CA

  • Min / @minrk / UC Berkeley / CA

  • Freek Pols / @freekpols / TU Delft / Netherlands

  • Giuliano Maciocci / @Gman0909 / 2i2c /UK

  • Faith Uniter N./ @afterthesegfault/ NAILUG ,F-Droid?/ Kenya

  • Jenny Wong (she/her) / @jnywong / 2i2c / Yorkshirrree, UK

  • Serena / @sbonaretti / Switzerland

  • April / @aprilmj / 2i2c / US East (Virginia)

  • Arielle / @Arielle-Bennett / The Alan Turing Institute / Boston MA

  • Brigitta Sipőcz / @bsipocz / Caltech

  • Fernando Pérez / @fperez / UC Berkeley / Berkeley, CA

  • Chris Holdgraf / @choldgraf / 2i2c / Berkeley, CA

  • Kelly Rowland / @kellyrowland / NERSC / Calif.

  • Ryan Lovett / @ryanlovett / UC Berkeley / Berkeley, CA

  • Angus Hollands / @agoose77 / 2i2c / Rugby, UK

Ice breaker 🧊⛏️#

What would you like our new community manager to know about you / your use of JupyterHub and/or Jupyter Book!

  • Min: I work all over JupyterHub, but I run fewer hubs than a lot of folks here. I work with Kirstie and Stefan at BIDS.

  • FP: Working in several smaller teams at our uni to increase open science/education through JB . Writing my portfolio now in JB.

  • Giuliano: I’m always interested in new use cases and existing pain points for JH

  • Jenny: Hey Serena! I would like you to know that I am super excited to meet you and support your endeavours as a community manager. I am a relatively new JH team member and would love to know how to open doors for others like me, if that’s something you are interested in

  • Erik: I deploy a JupyterHub, but have ran low on time to contribute overall, contributing sporadically. I have contributed for ~8 years or so, centered around JupyterHub in a Kubernetes context. I have appreciated your youtube videos Serena!!

  • April: I love other people who are interested in the social and human aspects of OSS, so looking forward to learning together <3 <3 <3

  • Angus: I’m stoked to see Serena is here! I just saw the mention on Zulip. I’m a JupyterBook maintainer, and JupyterHub contributor. I have a dog, if this is useful info!

  • Arielle: WELCOME SERENA! Excited to chat user workshops with you - I am also a member of the Turing Way which is a JupyterBook-based project :)

  • Brigitta: Welcome, it’s really exciting to have you here

  • Welcome Serena ! Excited to see you here!

  • fperez - Welcome Serena! So excited that you’re with us in a more official capacity :)

  • Ryan: i run some hubs for research in the Statistics and Economics Deprtments at Berkeley. I also participate in developing the undergraduate “DataHub” deployment.

Celebrations and shout-outs 🎉#

Add a note to celebrate the awesome work someone in the community has been doing!

Breakout room and agenda item suggestions#

Notes for Breakout room 1: MyST / JupyterBook#

Notes for Breakout room 2#

  • Learning from 2i2c migration, trouble with nginx-ingress

    • 2i2c’s learning in public has been hugely valuable for me to read, a great community value!<3

  • going to try traefik as hopefully simpler/smoother

  • Erik migrated to envoy-gateway this past weekend, deploy side-by-side

  • Gateway is more complex, envoy-gateway doesn’t define a default GatewayClass

  • Gateway == LoadBalancerService, by default

  • ListenerSet: https://gateway-api.sigs.k8s.io/guides/user-guides/listener-set/

    • may help with the tls disconnect

  • jupyterhub/binderhub#2107

    • more build traffic than I realized is triggered by crawlers and bots

Notes for Breakout room 3#

  • April, Jenny & Arielle watched Kelly & Yuvi review jupyterhub/jupyterhub-idle-culler#102

  • This a trial of live PR reviews at Collab Cafes - it was a success!

  • In the future might not be able to rely on contributors nudging maintainers, instead might need maintainers to be partially responsible for picking PRs to live review

  • Some things for future iterations - rotate maintainers, figure out what criteria make for a good live review, how do we approach people we don’t know as well to bring them in? Also, tell people these sessions are happening - useful skilling up for contributors looking to get more involved in reviewing as well!

  • The vibes were impeccable ✨

2026-05-05#

Check-in 💁#

  • Raniere Silva / @rgaiacs / GESIS / Germany

  • Kirstie Whitaker / @KirstieJane / UC Berkeley / Berkeley, CA

  • Min RK / @minrk / UC Berkeley / CA

  • Silas Santini / @pancakreport / UC Berkeley

  • Erik Sundell / @consideratio / Sundell open source

  • Jenny Wong / she, her / @jnywong / 2i2c

  • Kelly Rowland / @kellyrowland / NERSC / Calif., USA

  • Ryan Lovett / @ryanlovett / UC Berkeley

  • Chris Holdgraf / @choldgraf / 2i2c

  • Yuvi / @yuvipanda / 2i2c

  • Faith Uniter/ @afterthesegfault /NAILUG , FDroid / Nairobi Kenya

  • Michael Forbes / @mforbes / WSU / Washington, USA

  • Stéfan van der Walt / @stefanv / UC Berkeley + Scientific Python

  • Brigitta Sipőcz / @bsipocz / Caltech / Seattle

  • April Johnson / @aprilmj / 2i2c / Virginia

  • Harold Campbell / @haroldcampbell / 2i2c / Jamaica

Ice breaker 🧊⛏️#

Where are you looking forward to travel to in the summer (or winter if you’re in the southern hemisphere)

  • Raniere is visiting Beijing for the first time. <3

  • Min: Convict Lake (near Mammoth, eastern Sierra Nevada)

    • Stéfan: this is one of my favorite hikes in the area!

  • Erik: Fountainebleau forests with bouldering stones, south of Paris Jenny: i love this place! Sweeet

  • Kelly: Vancouver over Memorial Day weekend

  • Ryan: Cruise to Mexico

  • Yuvi: Finally able to travel to India after 18 months due to visa restrictions :(

  • Silas: New Orleans <– Chris lived there for 5 years lemme know if you want recommendations :-)

  • Brigitta - France for travel and tbd vacation with family somewhere in the Alps, otherwise being out in the mountains here as much as I can

Celebrations and shout-outs 🎉#

Breakout room and agenda item suggestions#

Notes for Breakout room 2: JupyterHub roadmap check-in#

  • Git repository at jupyterhub/roadmap

  • Yuvi: how to have people engage with the roadmap?

  • Jenny: not everyone attended the roadmap workshop. We have a separate repo.

  • Yuvi: for big chunk of work, have a description of the problem and the solution that we want. Not all new features need to go into the roadmap. Roadmap fails because it is not used by the community.

  • Use of jupyterhub-contrib

  • How do we make sure “issues” progress?

  • assign issues to shepherds?

  • have a shepherd for each issue to keep an eye on progress and activity?

  • Need clearer transition from ‘proposed’ to ‘accepted’

Myst#

  • Jupyter Book prio board: jupyter-book/projects#1

  • Demystifying myst workshop: jupyter-book/mystmd#2846

  • Path forward for integrating in Jupyter

    • Jupytext engagement

    • JupyterLab extension

  • Need a meta-issue to track AnyWidget vision

    • Examples:

      • Communicating between widgets

      • Give widgets access to app state

  • Cleaner Jupytext + mystmd integration

  • Better way to aggregate paper cut / experience reports

  • Stéfan: looking today at

  • Potential fundraising ideas:

    • I suspect we could motivate the role that MyST could play in AI-native science, and loop in a bunch of paper cut-style things in the process

      • Structured data, clearly defined narrative, interact with notebooks easily

    • Is there an advantage to having a CLI?

      • Enhancing the CLI with functionality that would allow AI skills to function (look up references, modify metadata, calculate diffs, output provenance tracking, etc.) — or do we simply write skills that understand myst markdown, the AST, and myst.yml?

    • Improving MyST Spec as source of truth for agent building and extraction (source of truth)

    • LSP / editor integrations to improve agent friendliness

    • Interoperable document AST (based on myst AST?): https://oxa.dev/

2026-04-21#

Check-in 💁#

  • Raniere Silva / @rgaiacs / GESIS / Germany

  • Kirstie Whitaker / @KirstieJane / UC Berkeley / Berkeley, CA

  • Ryan Lovett / @ryanlovett / UC Berkeley / Berkeley

  • Silas Santini / @pancakereport / UC Berkeley

  • Michael Forbes / @mforbes / WSU / Washington State

  • Harold Campbell / @haroldcampbell / 2i2c / Jamaica

  • Paolo Marzolo / @pollomarzo

  • Faith Uniter/@afterthesegfault/NAILUG/ Nairobi, Kenya

  • Stéfan van der Walt / @stefanv

  • Arielle Bennett / @Arielle-Bennett / Alan Turing Institute / Boston, USA

  • Angus Hollands / @agoose77 / 2i2c / Rugby, UK

  • Chris Holdgraf / @choldgraf / 2i2c / Oakland, CA

  • Freek Pols / @freekpols / TU Delft / Netherlands

  • Brigitta Sipőcz / @bsipocz / Caltech / Seattle

Ice breaker 🧊⛏️#

  • Ryan: Lottery winner press conference 🤣

  • Kirstie: FOSDEM or MozFest

  • Raniere: Python Brasil https://python.org.br/

  • Michael: Simons Workshop on Turbulence

  • Arielle: Collaborations Workshop in person; FOSDEM & FOSS backstage

  • Faith: FOSDEM or ATO

  • Angus: FOSDEM / scipy

Celebrations and shout-outs 🎉#

Add a note to celebrate the awesome work someone in the community has been doing!

Breakout room and agenda item suggestions#

  • Discussion of possible conference submissions: RSEcon, Compute!, and others

    • SciPy would be good but too late to submit for this year

  • Faith has submitted to PyCon Africa!

  • Key questions: topics of talks, why would we be submitting talks, what would we want to get out of it, how would this be funded?

  • Topics/why: showcase tech advances/features (e.g. Yuvi’s talk/demo at JupyterCon); showcase the work we’ve done on the roadmapping and community; explicit deployer/operator outreach similar to our user workshop at JupyterCon

  • What do we want to get out of it? Increase people on the contributor pathway; better discussions and insights into users; supporters or sponsors for JH or mybinder.org

  • Other ways we could look to achieve these outside of conferences? e.g. support from Linux?

  • Talked about having themes for conference talks for a year e.g. roadmapping; tech advances; community work

Notes for Breakout room 2#

  • Quick question: MyST for Blog Website (@choldgraf)

  • Question in discord about siunitx / physics packages. Would be nice to see a path through doing things like this.

  • jupyter-book/projects#1

  • jupyter-book/mystmd#2802

  • Two more a11y issues:

  • How are releases made?

    • as needed

    • interested in semi-regular (monthly, twice a month) release schedule to lower the scaries

    • Silas: would like another release by May 4

  • Ping discord for review of PRs

  • myst-theme repo could use axe checker in CI

    • some failures is expected; hope that number doesn’t go up

    • Keep track of failure count over time.

  • myst-theme should attach zip artifacts in PRs

    • would be good to document this workflow for authors who would use this flow

2026-04-07#

Check-in 💁#

  • Min (he/him) / @minrk / UC Berkeley (BIDS) / California

  • Kirstie (she/her) / @KirstieJane / UC Berkeley (BIDS) / California

  • Chris (he/him) / @choldgraf / 2i2c / California

  • Arielle (she/her) / @Arielle-Bennett / Turing Institute / Boston, MA

  • Raniere Silva / @rgaiacs / GESIS / Cologne, Germany

  • Kelly (she/they) / @kellyrowland / NERSC / Calif., USA

  • Silas Santini (they/them) / @pancakereport / UC Berkeley

  • Srihari Thyag (he/him) / @Haleshot / TN, IND

  • April (she/they / @aprilmj / 2i2c / Virginia

  • Harold / @haroldC / 2i2c/ Jamaica

  • Franklin (he/him) / @fwkoch / Curvenote / Calgary, AB, Canada

  • Jenny (she/her) / @jnywong / 2i2c / Yoo-kay

  • Ryan C. Cooper (he/him/his) / @cooperrc / UConn / Connecticut, USA

  • Angus Hollands (he/him) / @agoose77 / 2i2c / Rugby, UK

  • Yuvi / @yuvipanda / 2i2c / Oakland, CA, USA

  • Jim (he/him) / @JimMadge / Turing Institute / Exeter, UK

  • Faith Uniter( she /her) @afterthesegfault / NAILUG / Nairobi, Kenya

  • (after the fact) Ryan Lovett (he/him) / @ryanlovett / UC Berkeley

Ice breaker 🧊⛏️#

What is your favourite memory of Eid-al-Fitr (end of Ramadan) or Easter or start of Fall (in south hemisphere) or start of Spring (in north hemisphere)?

  • tulips blooming!

  • Cadbury’s Mini Eggs yummmmm

    • crunchy shell >>> +10000

    • Thoughts and comments on the changing of their taste

    • Not as good as they used to be but UK mini eggs >>> US mini eggs

  • Raniere enjoyed painting eggs during Easter for the first time when was 26 years old.

  • Seeing the flowers come through one by one

  • Nearly driving off the road looking at street trees in bloom

  • “So called heatwaves in the USA” aka not freezing cold

  • Its road+MTB cycling season in New England again!!

    • also of note: loved to have peep-jousting tournaments with my brother and a microwave

  • After my 1 week trip to Berkeley California, I slept a BUNCH and then painted my living room in “Tuscan Terracotta”

Celebrations and shout-outs 🎉#

Notes for Breakout room 1: JupyterHub Roadmapping#

  • jupyterhub/roadmap#issues

  • Kelly to add endorsement to Improving support for OIDC across JupyterHub #5

  • Kelly: open draft PR with batchspawner slurm improvements

  • possible hub roadmap item - acquire maintainer for batchspawner?

    • general question about what to do with retired/unmaintained repos in the project - jupyterhub-contrib as a possible option

  • Who would be a good maintainer for batchspawner?

    • Active user, with technical chops to support

  • Could we reach out to the wider community to solicit maintainers? Or the user community?

  • Who is using batchspawner with slurm vs other schedulers (e.g., torque)? How would we find out?

    • known knowns are UC Berkeley and NERSC currently using batchspawner with slurm

    • Ryan’s addition’s several hours after the meeting:

      • UC Berkeley’s Savio cluster doesn’t run JupyterHub. They use Open OnDemand.

      • I deploy JupyterHub on Slurm for the Statistics and Economics departments and we use batchspawner.

  • should we be focusing on adding more issues or moving to a human-oriented roadmap page like 2i2c has?

  • Things to add to roadmap:

  • Work through one item to completion

    • Candidate: Harbor since it’s almost done

  • Open source roadmap is not about constraining action, helps with “I have time, but I don’t know what to do”

  • Now we have enough information on the roadmap that issues list isn’t enough

    • first version: github project board

  • Project board: jupyterhub/projects#5

    • Will create issue on team-compass to discuss project board

Notes for Breakout room 2: A11y#

Notes for JupyterBook.pub#

Thank you for taking notes to allow members of our community to catch up if they weren’t able to make the meeting time.

  • Security boundary

    • Long run — tease out the implementation for building in a self-contained file to ease path towards containerisation.

  • Angus will restore support for custom themes for source builds still via two phase build.

    • Then merge this.

  • Yuvi / Angus will collaborate on the JupyterHub PR.

  • Yuvi / Angus will coordinate on security sandboxing and also on locking.

Notes for Breakout room 4: Jupyter Book roadmapping#

Thank you for taking notes to allow members of our community to catch up if they weren’t able to make the meeting time.

  • This was the first time we attempted to walk through the roadmap at this meeting

  • Chris did a little intro of the current project boards set up and intentions, but the rest of the room was full of lurkers/observers, so we couldn’t accomplish the goal of board review (which is alignment among contributors)

  • We all decided to go find other rooms :)\

2026-03-17#

Check-in 💁#

  • Kirstie Whitaker (she/her) / @KirstieJane / Berkeley Institute for Data Science / Berkeley, CA

  • Min RK (he/him) / @minrk / Berkeley Institute for Data Science / CA

  • Chris Holdgraf (he/him) / @choldgraf / 2i2c.org / California

  • Raniere Silva / @rgaiacs / GESIS / Germany

  • R Ely / @ohrely / Bloomberg / CA

  • Erik Sundell / @consideratio / Sundell open source /

  • Stéfan van der Walt / @stefanv / BIDS + Scientific Python

Ice breaker 🧊⛏️#

Favourite flavor of ice cream? And favorite place to eat it!

Celebrations and shout-outs 🎉#

Add a note to celebrate the awesome work someone in the community has been doing!

Notes for Breakout room 1: mybinder#

  • We will change the Harbor configuration to use a quota of 10TB

  • Harbor is configured with retention rules that we are still assessing as it takes time for Harbor to “flat” in storage thanks to garbage collector

  • We will explore the option to sync the container storage between OVH and Heztner and revisit if it become too expensive given egress costs

  • The “old” 70TB s3 bucket will be deleted soon.

Notes for Breakout room 2: Jupyter Book#

  • It can be hard to figure out how to do something that isn’t already clear

  • Accessibility -

    • Jupyter Book 2 - uses MyST - uses ReDux HTML framework - better accessibility than JB1

    • How do we check for the math accessibility

  • Proposed enhancement - can we make sure to add alt text to ….. [HELP - what?]

    • Can we add alt text to tables [More details?]

  • Need to figure out which customisations are “MyST’s problem” vs work for the person / team _doing_ the customisation!

    • The MyST problem is to ensure there are pathways like AnyWidget so that when users run into problems that are outside of the MyST perspective, they have a path (hack) to get what they need done.

  • Do we want to make it easier to load in custom HTML?

    • How do we make it safe to include?

  • Conversion to LaTeX looks pretty smooth.

    • What is reassuring about this is that MyST -> LaTeX is pretty straightforward and can be easily customized. From this “checkpoint” one can confidently generate accessible PDFs using standard procedures.

  • This is a relevant issue: jupyter-book/mystmd#2743

  • Question about how to configure internal things like the command options used when running latexmk (see https://github.com/jupyter-book/mystmd/issues/1855)). Conclusion: no specific place for this right now. Discuss on Discord.

Notes for Breakout room 3: Jupyter Health#

  • Case study: JupyterHealth wants to support 2i2c’s upstream contributions to the tools that JupyterHealth depends on.

    • What could 2i2c provide to Jupyter Health project. And how Jupyter Health could grow from part of JupyterHub into its own project?

  • What is needed to sustain projects and communities outside of contractual work?

  • Foundational contributions to the tools that JupyterHealth depends on

    • What are the foundational where we?

2026-02-17#

Check-in 💁#

  • Kirstie Whitaker (she/her) / KirstieJane / Berkeley Institute for Data Science, UC Berkeley / Berkeley, CA

  • Min (he/him) / minrk / BIDS / California

  • Simon (he/him) / manics / University of Dundee / 🏴󠁧󠁢󠁳󠁣󠁴󠁿

  • Ryan (h/him) / ryanlovett / UC Berkeley

  • Brigitta Sipőcz / @bsipocz / Caltech / Seattle, WA

  • Kelly Rowland (she/they) / @kellyrowland / NERSC / SF Bay Area, CA, USA

  • Jenny Wong (she/her) / 2i2c / Yorkshiree UK

  • Yuvi / 2i2c / Oakland, CA, USA

  • Arielle Bennett / @Arielle-Bennett / The Alan Turing Institute / Boston MA

  • Chris Holdgraf / @choldgraf / 2i2c / California

  • Angus / @agoose77 / 2i2c / UK

  • Jonathan Guinegagne / @JGuinegagne / AWS / New York, NY, USA

Ice breaker 🧊⛏️#

Favourite Winter Olympic sport? (To watch or participate in!)

  • Kirstie: curling I think! Or snowboard cross?

  • Simon: also curling, surprisingly strategic based on random YouTube/Instagram Shorts

  • Ryan: Luge? Bobsled?

  • Yuvi: Too cold for me to watch

  • Angus: LUUUGE. Also skating.\

  • Jenny: I have not been following, but seen random clips that look funny – such as a few photo finishes with athletes in compromising(?) looking positions. Such a Jenny coded comment

  • Brigitta: I’m not really watching the games, but already planning a curling outing with my alpine club (and can’t wait to get more snow)

  • Arielle: Snowboarding (I am a very mediocre snowboarder) & ICE HOCKEY - USA vs Canada final for the women’s gold! Records being broken all over the place!

  • Kelly - all the things! big olympics fan generally, ice hockey fan outside of the events

Celebrations and shout-outs 🎉#

Add a note to celebrate the awesome work someone in the community has been doing!

  • Jenny Wong added to the team!!

    • Congratulations Jenny!! !!!!

Breakout room and agenda item suggestions#

  • Roadmapping workshop planning - agenda, activities, what work do we need to do before February 2026?

  • Yuvi: jupyterbook.pub and binderlite could be the same thing

    • @agoose77

  • Yuvi: Integrating repoproviders into repo2docker / binderhub / nbgitpuller

    • @agoose77

    • jnywong: also interested but i feel that would be too 2i2c heavy

  • CH: Anybody want to brainstorm where a JupyterHub landing page should be?

    • Yes at a cafe in March pls!

  • Angus: Thinking about anywidgets?

Notes for Breakout room: Jupyter Hub Roadmapping#

  • What are we worried about?

    • Kirstie: Small group in Europe - want to make sure that team are well set up for success

      • Quantstack COO Matthias attending - what opportunity do we have to engage with him?

        • Is he the one who was interested in TREs?

          • Yes - potentially - although we should confirm that….

            • Suggestion from Kirstie: Use Europe session 1 to focus on TREs, save the Roadmap/Goals for the first synchronous session

    • Arielle: Best way to make use of synchronous sessions

    • How to convey ideas between groups (4 groups - local/remote, EU/US)

  • What do we want to achieve?

    • establish goals for roadmap

      • What do we mean by a “goal”?

    • discuss current/future plan options, start prioritizing

  • We need to ensure

  • Jenny is not attending but happy to put cycles into help with any last min prep

  • CH: Which of the following is the best example of our goals?

    • “expansive” (e.g. brainstorm lots of ideas)

    • “contractive” (e.g., turn a lot of ideas into a prioritized list)

    • “Process-oriented” (e.g., meta conversations about what a roadmap means to us and how we’ll use it)

    • Answer:

      • Day 1: Expansive. Strategic analysis w/ an external or product focus. Where does JupyterHub fit in the landscape. When we think of the purpose of the broader Jupyter project, are we aligning ourselves with that purpose? SWOT style analysis.

      • Day 2: Expansive. Strategic analysis w/ an internal or team focus. How does the roadmap serve the community? This is NOT “how do we manage the roadmap?” “how is it structured?” “who decides what goes on the roadmap?” etc.

      • Content differences:

        • EU session - trusted research environments

        • California session - educational researchers

  • Existing resources

  • Asynchronous work:

    • Ideas/activities for prioritisation

  • Each day: 3h Europe session, 3h joint session, 3h US session

  • Europe: recommend a deep dive into scoping the TRE use cases

  • Day 1 Synchronous work (3 hours):

    • Introductions - based on the presentation Min has prepped

      • 30 mins?

    • Mapping the landscape - where is JupyterHub active

      • Breakouts: 30 mins

      • Summaries: 15 mins

    • Tightening up the themes of the landscape

      • Breakouts?

      • Shared discussion

      • Is there anything we would NOT do based on this

  • Day 1 outcomes:

    • Short page that articulates what purposes the JupyterHub community is looking to serve

  • Day 2: how can the roadmap serve the community?

    • Developing themes: TREs, education,

      • Who will be keeping an eye on parts of the project?

  • Shared statement to promote whats happening is probably something to happen after the workshop - might be difficult to achieve consensus

    • Its an outcome of the workshop but not completed during the workshop

    • Aimed at exec council

Notes for Breakout room: JupyterLite / BinderLite / Jupyter Book#

  • Yuvi was at the Project Pythia community meeting today, where the conversation about using jupyterbook.pub for PR previews came up.

    • Brigitta talks about CircleCI usage vs Netlify. A possible goal to ditch circleCI?

    • We should create an action to make this trivial!

    • JB needs to support do-not-execute.

    • Pass in JUPYTER_SERVER env vars like we do in?

      • Unbounded execution not ideal from cost perspective.

      • Bring your own execute.

    • Build / API tokens?

      • Not unless we have to — complexity/security plus ecosystem vibes

  • Caltech-IPAC/irsa-tutorials#253 was made as we try to figure out if what else is needed to support executed previews on jupyterbook.pub

  • Yuvi demos building binderlite on top of jupyterbook.pub foundations.

    • Uses emscripten-forge

2026-02-03#

Check-in 💁#

  • Kirstie Whitaker (she/her) / @KirstieJane / UC Berkeley / Berkeley, CA

  • Brigitta Sipőcz (she/her) / @bsipocz / Caltech / Seattle, WA

  • Michael McNeil Forbes / @mforbes / Washington State University / Pullman, WA

  • Raniere Silva / @rgaiacs / GESIS / Germany

  • Paolo Marzolo / @pollomarzo / politechnic of milan / milan, IT

  • Angus Hollands / @agoose77 / 2i2c / United Kingdom

  • Min / @minrk / UC Berkeley / Berkeley, CA

  • Yuvi / @yuvipanda / 2i2c / “United” States

  • Jenny (she/her) / @jnywong / 2i2c / UK

  • April (she/they) / @aprilmj / 2i2c / Virginia, US

  • Stéfan van der Walt / @stefanv / BIDS + Scientific Python / Truckee, CA

  • Silas Santini / @pancakereport / UC Berkeley

Ice breaker 🧊⛏️#

What costume are you planning for Carnival or Halloween?

Celebrations and shout-outs 🎉#

  • Kirstie: Jupyter Health team having a team retreat - considering how we can work more in the open!

  • Kirstie: Big docs overview from Chris!!

  • Chris: Thanks to Raneire for posting about the collaboration cafe on LinkedIn

  • Chris: Thanks to Angus for pushing forward that MyST outputs AST PR a bit more (I know we still have a ways to go)

  • Yuvi: Thanks to Min for getting BIDS sponsored OVH node on mybinder.org fully done

  • Yuvi: Thanks to Jenny for getting a significant UX change into nbgitpuller that will make a lot of instructors & students’ lives better 🙌

Breakout room and agenda item suggestions#

Notes for JupyterHub Roadmapping workshop#

  • jupyterhub/team-compass#805

  • The agenda (only has topics for the Europe time zone so far) for reference: https://docs.google.com/document/d/1k0RcJR_g1WgJzTyU3a8paQUg6H_hlTt0n8fSnZU7Oa4/edit?tab=t.0#heading=h.395htu1urps

  • Question from Chris - what outputs from the roadmap workshop will we share with other subprojects?

  • Linux Foundation would like a more detailed/finalised agenda before this gets published to the website

    • How do we “pass the baton” between the EU and US sync sessions?

    • The problem to solve: what’s a priority and how to communicate that to someone outside the project?

    • Roadmap purpose: welcome new people and use it as a way to prioritise and have the power to say no to extraneous things

    • More intentional contributions than reactive

  • Kirstie’s proposal: Day 1: get everything on the table, day 2: narrow and focus

  • Min: would love a way to make the roadmap a continuous living document

    • failure mode is writing a roadmap that doesn’t get used

  • Consider the participants in the room, how do we get the most out of them

  • Day 1 -> easy to see that there’s too much to do, Day 2 -> impetus for people to design how we are going to prioritise, not what

  • Nice to have regular breaks so that people have opportunities to switch to different rooms/breakout topics

  • Diamond framework: diverge in the async sessions, then converge in the sync sessions

  • day 0: slide deck for intros in advance so we get a sense of who the participants are and what they care about https://docs.google.com/presentation/d/1RRPqdKtNWmK1sBrYzKB0tnUJfpf5avw8s8NxlmkjWrQ

  • day 1 async: ideas for what we should be doing (divergent)

  • day 1 sync: align on roadmap goals (converge, more structured)

  • day 2: bring it together, synthesis, how do we prioritize

  • a more sustainable way forward could be that people can come away feeling like they are stewards (of all or) different parts of the JH ecosystem

  • JH is not a monolith; we can subset and have smaller teams, e.g. TRE, binder

  • Challenge for the project is not to have too many overlapping team members so that they are not overburdened to keep track of more moving parts: 7 priority lists are harder to digest than 1

  • For the workshop, it might be a good exercise for subteams to polish what they want in the roadmap

  • Day 1 spring cleaning gives us signals as to what people care about

  • Previous failure modes for roadmapping:

    • no process for updating them

    • no bearing on work that was happening – felt more like a blog post on what was happening now, and not for guiding work and keeping communication open

    • not designed to serve the purpose of where to prioritise time, more like a snapshot of the current work/retro

    • previous roadmaps could be intimidating to newcomers? Tradeoff between a roadmap audience for newcomers/seasoned maintainers

  • EU async timezones are harder to structure since they spark the discussion that is passed onto the US timezone – setting expectations, lightning talks

  • US timezone can refine and pass back proposals to EU

Notes for Breakout room JupyterBook / MyST a11y#

None.

Notes for Breakout room JupyterBook / MyST AST#

Thank you for taking notes to allow members of our community to catch up if they weren’t able to make the meeting time.

jupyter-book/jupyter-book#2578

toc:
    - file: parent
      children:
          - file: child

parent.md:

# Title (should be 5.0)

## Heading 1 (should become 5.0.1)

child.md:

# Title (should become 5.1)

## Heading 1 (should become 5.1.1)

2026-01-20#

Check-in 💁#

  • Min (he/him) / @minrk / UC Berkeley / California

  • Kirstie (she/her) / @KirstieJane / UC Berkeley / Berkeley, CA, USA

  • Tyler (he/him) / @tylere / VorGeo / Los Altos, CA, USA

  • Freek / @freekpols / Delft University of Technology

  • April (she/they) / @aprilmj / 2i2c

  • Raniere Silva / @rgaiacs / GESIS

  • Kelly Rowland (she/they) / @kellyrowland / NERSC / Bay Area CA

  • Ryan Lovett / @ryanlovett / UC Berkeley / California

  • Angus Hollands / @agoose77 / 2i2c & JB / Rugby, UK

  • Simon Li / @manics / University of Dundee

  • Jenny Wong (she/her) / @jnywong / 2i2c / Yorkshire UK

  • Arielle Bennett (she/her) / @Arielle-Bennett / The Alan Turing Institute / Boston, MA

  • Yuvi / @yuvipanda / 2i2c

  • Brigitta Sipőcz / @bsipocz / Caltech / Seattle, WA

  • Chris Holdgraf / @choldgraf / 2i2c

  • Rowan Cockett / @rowanc1 / Curvenote

Ice breaker 🧊⛏️#

What have you already completed in 2026? (Can be professional, personal, related to Jupyter or not!)

  • Walking without stick

  • [Kirstie] Finished Friday Night Lights and The Warmth of Other Suns this weekend! Both pretty big books and absolutely fascinating for an immigrant to the USA!

  • [min] Deployed a JupyterHub with Gateway API (envoy-gateway), migrated away from nginx-ingress

    • YAAAAS let’s go! (Angus)

  • [Angus] I have read three *dreadful, awful, bottom tier *books

    • do tell tho (Kelly)

    • Tell meeee which ones 👀

  • Raniere] Watched the first season of “The Big Bang Theory” in German.

  • [jenny] making some nbgitpuller improvements jupyterhub/nbgitpuller#383

    • also, i went to paris to extract a cat to the uk last week on an EPIC road trip

  • I build a prototype {embed} role for MyST.

  • [Ryan] Improved my homeassistant deployment. Chopped up a tree that fell down in my yard.

  • [Arielle] The Bright Sword, Katabasis, and StoryLand (I have Opinions)

  • [Brigitta] have not completed anything :(

Celebrations and shout-outs 🎉#

  • Got new inputs for gallery :D 🖼️ Gallery of Jupyter Books - Jupyter Book

  • We have people signed up for the Roadmapping Workshop! Shaping up to be a really interesting event in Feb.

  • Chris’ new plugin, choldgraf/myst-substitutions

  • Agah worked on jupyter-book/mystmd#2413, which th Jupyter Book project finally merged! Thanks Agah!!

Notes for JupyterHub Roadmapping workshop#

  • jupyterhub/team-compass#805

  • there are 3 big timeslots we would like to plan an agenda for

  • when do we want breakout rooms? when do we want everyone in a sync room together?

  • what’s the purpose for maintainers and deployers as a source of info?

  • what should we populate the roadmap with?

  • discussing the purpose of the workshop

    • this feels like we should involve everyone

    • defines the parameters of the workshop, e.g. what is a roadmap

  • 2 days: 1. what is a roadmap, 2. populate the roadmap

    • theme 1: priorities?

    • theme 2: process?

  • folks involved with TREs tend to be based in EU

  • ask people to prep priorities in advance and host a discussion group during the sync time? i.e. unconference style

    • possibly some project management here, how to break tasks down

  • or traditional approach: put them on a post-it, rank and then order?

  • resources and capacity should inform what should be on the roadmap? We are not top-down, we are a collection of people with needs, goals, and varied availability

  • it shouldn’t be a to-do list, could have “side-quests” in terms of communicating priorities

  • roadmap is not a linear path

    • what questions do we want to answer?

    • deployers might want to know next release

    • contributors: where can i direct my time?

  • circulate who’s who ahead of the meeting, that might help identify personas and themes for roadmap

  • Day 1: purpose of the roadmap, proposals on how we approach that, e.g. how to keep it up to date, something that is useful for folks to look up easily,

  • primary goal is to come out with a process that we can use for the community at large to continually engage with the roadmap and gather inputs, and design it in a way that fits people’s interests and priorities

    • e.g. if folks are interested in security, then how does this fit with a shared roadmap?

    • useful to collect experiences from a project planning perspective and avoid common pitfalls

    • could do some skills matrix mapping ahead of time

  • follow-up work after the workshop

  • personal experiences with roadmaps

    • a public roadmap is not a project management tool, but a tool for helping to coalesce communities and stakeholders together to drive forward progress

    • if you have a team with resources, then this can be useful, but if you have a fragmented, overstretched team, then the roadmap tends to get ignored

  • aspirational or practical?

    • difficulty * effort * priority

    • e.g. lots of effort but important tends to get left behind, idea for multiple replicas for hubs has been around for a while, but has not got traction

    • how to make this more visible for potential contributors so that they feel like they can tackle roadmap items?

  • outcome of roadmap could help draw in funding from grants to buy people’s time

  • process could be to sort items into priority and resource buckets

    • how do we present and update this information easily?

    • will the community manager be able to help here?

  • Next steps after this café:

    • take the icebreaker slides away and prep them

    • invite some registered participants to future collab cafés to help discuss agenda

Notes for Jupyter Book / MyST release#

Notes for Community Manager recruitment#

Notes for What do you want from the Jupyter Executive Council!?#

No notes.

2026-01-06#

Check-in 💁#

  • Name (and pronouns if you’d like) / GitHub handle / affiliation / geographic location

  • Raniere Silva / rgaiacs / GESIS / Cologne, Germany

  • Kirstie Whitaker / @KirstieJane /

  • Jenny Wong (she/her) / @jnywong / 2i2c / Yorkshire, UK

  • Brigitta Sipőcz / @bsipocz / Caltech / Seattle, USA

  • Stéfan van der Walt / @stefanv / BIDS + Scientific Python / Truckee, CA

  • April Johnson (she/they) / @aprilmj / 2i2c / Virginia, US

  • Freek Pols /@freekpols/ delft

  • Angus Hollands / @agoose77 / 2i2c & Jupyter Book

  • Robert Lanzafame rlanzafame GEI Consulants Oakland California USA

  • Thierry Parmentelat - Inria & École des Mines - France

  • Yuvi / @yuvipanda / 2i2c / Oakland, CA

  • Silas Santini/ @pancakereport / UC Berkeley

Ice breaker 🧊⛏️#

What do you hope to achieve in 2026? (Can be professional, personal, related to Jupyter or not!)

  • Raniere wants to reach level B1 on German language skills.

  • Angus wants to read more books, and play more guitar. Also, more snek jokes. ALSO JB featurreeees

  • Kirstie wants to figure out how to connect multiple conversation workspaces and promote events in a coordinated way!

  • April sets intentions in September & wants to focus on making and spreading joy

  • Jenny wants to survive and go travel and chill by a beach somewhere

  • freek, finally get some papers out.

Celebrations and shout-outs 🎉#

  • Kirstie: BIDS is now part of the MyBinder Federation!

  • Angus: Jupyter Book - nearly at parity with JB1 on output handling!

  • latest page Update plugin

  • luuk gets the wysiwyg ready

  • Yuvi wants to demo jupyterbook.live or jupyterbook.pub or jupyterbookbinder or unnamed thing (not sure where this goes)

Notes for Jupyter Book#

  • Injecting Javascript into sites

  • Accessibility

    • Had a pass early on in the project

    • Still needs work, opportunity for someone to take a lead on

    • There are a few low-hanging issues, such as color contrast

      • Cell errors don’t meet color contrast requirements

      • Scroller issue ?

    • See also Silas’s notes below

  • Python Theme server proof-of-concept

    • Show how a “theme” works

    • Consume AST and files and render output

    • Alternative for low-level rendering without JavaScript stack

  • MEP status update

  • Outputs work update:

  • Universal AST - https://discord.com/channels/1083088970059096114/1457711013532139551/1457711030871527664

  • Extensible MyST Exporters — https://hackmd.io/h8NTuPFBQImhCKKELOhx2g

  • 2026 big picture items

    • What is the most useful way for the community to help?

    • Execution initiative: jupyter-book/mystmd#2019

    • We’re still working on a good strategy for prioritizing and executing on issues

  • Reminder: weekly work sessions (see https://compass.jupyterbook.org/contribute/#events))

    • Wed 9am Pacific, and feel free to organize other sessions

  • Yuvi demo: yuvipanda/jupyterbook.pub

    • enter repo (does not need to be GH) and static files for a book are deployed (magic?)

    • yuvi has 5 hrs into this tool and wants input from JB users to know where to take it further

    • ideas:

      • from demo chat: helping with PR review, preserving older/alternative versions of books, …

      • ability to edit/modify static files after build?

      • regarding execution of source code during book build (e.g., executing cells of a nb); is that possible, and if so how; if it needs a handoff to some compute engine how would that work?

      • use case for a user training: participants edit content, use tool, view book. should be straightforward way to see all of the “version” (perhaps by commit for new users?) and manage how many are preserved? Perhaps a sort of “refresh” button type of feature that jumps to the build from the most recent commit from a repo

  • freek/luuk will be working on wysiwyg tool coming weeks

  • Silas’s Concerns RE Accessibility: color contrast for erroring cell output jupyter-book/myst-theme#744 and scrollable regions (e.g. wide code cells) jupyter-book/myst-theme#743. Need help figuring out where to get started, but should be able to make the fixes easily.

  • Also interested in learning more about interactive visualizations. Discussed kernel-based ones (ipywidgets), non-kernel based ones (plotly, bokeh), iframes (embed a pre-baked visualization). also anywidget and using JS-directly (not yet available, but desired)

Notes for Community Manager recruitment#

  • Proposal: jupyter-governance/funding-proposals#6

    • Funded at 90k

  • We need to know what success looks like, so that we can assess whether someone is well suited for the role

    • We’re likely to get N>1 good applications!

  • Paid take home assignments are a great way to actually see someone’s work

  • Can we also collect examples of where they have done similar work before

    • Both as stories (behavioralish interviewing process) and as a portfolio of past work

  • We need to know who is going to be part of the hiring process

    • Everyone knows everyone else - and it is very likely that the best candidates WILL be known to us!!

    • Kirstie: FINAL decision maker, chair of the committee!

    • Yuvi: wants to be part of the process

    • Carol Willing: Yuvi would like her to be part

    • Chris Holdgraf: doesn’t need to be part, we think? (ask him) (similarly not Angus as 2i2c represented already through Yuvi)

    • Suggest one of Stefan or Rowan - ask them to nominate?’

    • Min and April: both would like to help design and run the process, to be fair, inclusive & make a good decision (neither needs to be a decision-maker)

  • What’s the recruiting pipeline and timeline

    • Steps

    • When do we want to start?

    • When do we hope to finish?

    • What tools do we use for recruiting (applicant tracking system, etc?)

  • How much time do we expect the person to spend on the project?

    • Can they just tell us how much time they’ll spend for $90k?

  • How senior is this role?

  • Are they doing strategy work? Are they writing reports? What percentage of their time will they spend on this effort?

  • Are they engaging with the community? Are they onboarding folks? Supporting newcomers participation?

  • We need the strategic thinker expertise - doesn’t need to be super tenured

  • Needs to wrangle technical leadership (senior, busy, cat-shaped)

  • Confident, skilled, trusted are qualities we need to look for

  • They need to set a direction AND go do work hands-on AND bring other people along with them - figure out what to do, do it, get others psyched to do it

  • How heavily do we want to weight knowing the JH/JB tech stack?

    • You need to understand it well enough that it’s not a barrier to setting direction

    • Needs to be empathetic to users and developers!

  • As trite as “fast learner” is, the person needs to very quickly learn enough of the power dynamics of the community, enough about the people and vibes, enough about the tech stack, and enough about user experiences to have a thoughtful, trusted - lovable even - voice

  • A skill that is required is being able to montior all the different connections and communications for the projects (discord, github, zulip etc etc) and be able to direct the questions / suggestions to the right place or person.

    • Very high trust and respect from maintainers required!

  • Actions:

    • By next collab cafe, Kirstie, April (and Min?) will draft a process

    • Yuvi will own creating a take home assignment (deadline 1-2 weeks after that)

On this page
Edit on GitHub
Show Source