2026 Open edX Conference Talk Topics Brainstorm

Discussions - Public
1

You will be assigned dedicated tickets on which to log time. Until then, feel free to ignore this thread.

This is the brainstorm thread for talks topics for the 2026 conference.

There will be three tasks - the current one for the brainstorm, a second one for voting and assignment, and a final one to write up the actual proposals. The whole team participates in brainstorming talk ideas and submitting them, regardless of who actually attends the conference.

Timeline

Current task

  • Submit three talk ideas.
  • Also comment on any issues you think may have resulted in us getting fewer acceptances last year, and any ideas you think might make talks more viable this year.

Next sprint

Weā€™ll vote on the talk ideas and assign them from the top rated ones.

First new sprint in January

We will write, review, and submit talks.

Getting ideas

You can read the same thread from previous years for ideas:

Avoid suggesting tutorials or ā€œHOWTOsā€ - these ideas are easy to get, but they take a lot of work and time to prepare afterwards, and usually have smaller attendance.

In general, try to focus on talks that our clients or prospects might attend - it helps a lot if itā€™s not only technical, but also focuses on outcomes for the learners and the instructors, or the community at large.

Think about what you have worked over the last year ā€“ what got you or our clients excited? The more you use materials you have already worked on, the easiest it will be to prepare the talk. Avoid new topics or ideas specifically that require development or research specifically for the talk.

Be opinionated in your talk title (think of what would make you click on an article title) - it should attract the attention, and make people curious about what you are going to say.

Information based on last years agenda:

AI was very heavy last year. It may have been a component in the majority of talks. My perception is that the excitement has cooled somewhat but I donā€™t think weā€™re in the trough of disillusionment just yet, so submitting talks with an AI bent is probably still a viable approach.

Another big one is showing off new product features weā€™ve developed and cool things you can do with them.

A third was talks that covered topics that are useful for MOOCs in particular-- Projects that deal with scale and data.

Consider the above categories first, but if you have something you think is especially interesting that doesnā€™t fit into this category, still submit it. We donā€™t want things to be too homogeneous, after all :slight_smile:

2 Likes
2

@Fox Thanks! Can we do only the brainstorm in the first iteration, as I was suggesting? Otherwise we often choose a talk based on a partial list, and we end up selecting & brainstorming at the same time. We didnā€™t get a lot of talks accepted lately so I think we should try to change the approach a bit.

I would focus only on the brainstorm, specifically aiming to come up with suggestions that would have more chances to be accepted than last yearā€™s batch. Worth including this in the brainstorming? And maybe we could even vote on which talks to submit first, then work our way from the most voted proposals to assign them afterwards?

1 Like
3

@antoviaque I think I understand, but Iā€™m confused on a few points:

  1. For clarity, you want to make the first task ONLY about brainstorming, right? So one full sprint for brainstorming?
  2. Then, selection would need to be put in the second sprint task? Or are we not doing the second task as currently planned, instead replacing it with a voting session?
  3. If weā€™re doing a voting session, does that mean that not everyone is going to submit proposals? Or is it that weā€™re using the voting session to rank them such that the least preferred options drop off the bottom of the list, but weā€™re all still submitting one?
  4. Does that mean submission of the proposals is moved to a third sprint? Note that it will definitely land in the holidays (end of December) but we know that submissions are opening soon (sometime in December, unspecified).

So you know, I did try to figure out what talks were accepted. However, of our talks the main takeaway was that ā€˜AI was more likely to be accepted, but we had so few talks accepted that itā€™s not clear how important that wasā€™ and I couldnā€™t find a persisting record of the schedule from last year.

However, I just remembered this morning that we used the Whova app last year, and Iā€™ve found that it does have the agenda, so Iā€™ll update the descriptions with patterns I find (Iā€™ve already found one-- and yes, it was AI.)

4

@Fox To answer you:

  1. Yes, including suggesting at least 3+ ideas of talks (for the part below to work) - and also brainstorming about the reasons why so few talks were accepted.
  2. Voting & assignment during that second sprint yes - maybe just a couple of days to vote at the beginning of the sprint, then assigning them by the end of the first week, and writing the proposal the second week?
  3. No, it means that instead of each person picking their own proposals (which is mostly what happened in previous years), we would pick 1 or 2 proposals each, starting from the top. So if your proposals arenā€™t among the top-voted, you need to pick another proposal which is.
  4. Itā€™s true that the 2nd sprint would be a bit packed. We could probably do it over 3 sprints and wait for January for writing the proposals - the advantage of starting early is that it gives us space to do that, so :+1:

And good to know there was a trend for AIā€¦ Ah, the technical fads! Imho we donā€™t need to aim specifically at submitting AI-related talks just to match the trend of the moment. But that can be good to have in mind.

5

Thanks, @antoviaque . That helps.

Iā€™ve updated the tickets Iā€™ve created, moving the writing portion to the first sprint in January, and making the brainstorming session all of next sprint.

Next week Iā€™ll make the intermediary tickets for voting and assignment.

1 Like
6

Here are my three talk ideas:

  1. How can I use AI on my Open edX Instance today? Navigating what tools and integrations are available and production-ready.
  2. Community Retrospective: The Post-Acquisition Transition (I still really want to do this talk. And I think people would love to hear it. Now that Iā€™ve written a blog post, it might be an easier sell)
  3. Working with Product: How the Product Proposal Process Makes Engineering Better

So, I updated the thread opener with what patterns I noticed for accepted talks. But aside from things we submitted which werenā€™t in pattern, I think another factor was the fact we were (are? It sounded last year like weā€™d be doing this moving forward, but it seems less likely with the event being hosted US-side) sharing the convention space with EMOOCS, which meant that talks were were split between the two major groups. While there was overlap, EMOOCS needed dedicated talk slots that we werenā€™t likely to be in the running for at all.

3 Likes
7

My ideas for talks:

  • The typesense benchmarking epic (BB-9937) was quite interesting and I feel there could be potential for an interesting talk. Some ideas that could be included:
    • importance of considering scale when developing for Open edX
    • Benchmarking retrospective, takeaways, and caveats - how it could be applied to other parts of the platform.
    • comparing apples to oranges: typesense vs meilisearch
  • Something about ā€œmodernā€ XBlock development? We have XBlocks developed in React now, itā€™s pretty convenient working with tutor mounts on devstacks now, and maybe other new improvements to development, testing, and general quality of life?
  • Tips and tricks for working with Tutor devstacks (eg. multiple local devstacks, quicker methods to get changes than re-launching, cleaning, Mise and virtual environments for devstack dependency management, etc.
3 Likes
8

My talk ideas:

  • About Content Libraries:

    • A workshop that demonstrates the functionalities of Content Libraries. Last year, Jill gave a talk on the basic functionalities of content libraries, and it was a success. Although there wasnā€™t enough time to show everything that could be done, people had many questions and wanted to know more about how it worked and try it out.
    • Content synchronization in Content libraries. It is a feature of Content Libraries that has been growing and becoming more complex over time. We can have two approaches:
      • A more technical one that explains what upstream links are and complex things like top-level parents.
      • A more user-friendly one that focuses on explaining the functionality.
    • Modulestore migrator: Explain how it works and how it can be used to migrate legacy libraries and courses to content libraries.

  • Optimistic update using React Query. With some examples within frontend-app-authoring. Itā€™s a very technical talk, but might fit better as a Lightning Talk.

Iā€™m not very familiar with the topic, but it could be: Migrating a legacy XBlock to React/XBlock V2 (?)

3 Likes
9

Talk Ideas:

  1. Staying Ahead of Digital Dishonesty: Safeguarding Online Assessments with a New Proctoring Tool
    • We submitted this proposal along with Elizabeth from ASU (she wrote the bulk of it, while @tecoholic and I made some additions and adjustments). The proposal is currently in product review.
  1. Open edX in Action: 20 Stories of Flexibility and Global Impact
    • We have designed about 15 (and counting) case studies for Axim describing how institutions around the world use Open edX. I thought of showcasing these stories to emphasize the platformā€™s flexibility and the inspiring ways itā€™s being used around the world.
  1. Empowering Admins: Edit, Translate, and Share Open edX Notifications with Your Course Teams
    • This is based on a proposal thatā€™s still in its very early stages. Although it has been presented to the TOC who shared a lot of feedback (which I havenā€™t incorporated yet), thereā€™s no guarantee the proposal will be accepted.
2 Likes
10

My talk ideas:

  1. Learning Beyond Courses: Shaping Open edXā€™s Future with Learning Paths
    How the new Modular Learning Paths foundation will help Open edX support stackable skills, flexible progression, and continuous learning. This is based on this proposal.

  2. Choosing the Right LMS for the Future: A Decision Framework for Universities, Governments & Enterprises
    This talk presents a practical decision framework for evaluating modern LMS platforms, including key considerations around flexibility, scalability, compliance, ownership, and integration based on our pillar post. The talk is based on our pillar post.

  3. Learning Paths: What Authors and Learners Can Look Forward To
    A forward-looking co-presentation with Edly exploring the future authoring and learner experiences for Learning Paths. This talk highlights the UX vision weā€™re building as a community, grounded in the proposals already underway (this is future work, and weā€™re not guaranteed it yet).

3 Likes
11

My talk ideas:

  1. Importance of non-linear learning: A talk about why linear course structures often fall short, and how non-linear, choice-based learning can dramatically improve engagement and retention. Using the Branching Scenario XBlock as a concrete example, it highlights the power of giving learners meaningful choices and makes the case for why Open edX should support non-linear learning natively.
  2. Microlearning tools to boost engagement: A talk focused on how simple, lightweight interactions, especially retrieval-practice tools like flashcards, can significantly boost learner engagement in courses.
  3. Learning through conversation: A talk centered on the emerging value of conversational practice as a learning approach, using the Coaching AI XBlock as an example. It explains why guided, reflective, multi-turn dialogue is effective for developing reasoning and soft skills, and how AI-driven coaching can be leveraged to bring conversational learning to Open edX.
2 Likes
12

Here are my talk ideas:

The Upgrade Dilemma: Lessons from Real-World Open edX Migrations

Upgrading Open edX is more than a technical task, itā€™s a strategic challenge that affects clients, instructors and the broader learning ecosystem. In this session, weā€™ll dive into the real-world complexities of upgrades, drawing on our recent migration and platform update for Penn State. Attendees will get an inside look at the technical and organizational hurdles that make upgrades such a hot topic in the community, from managing customizations and integrations to balancing risk and resource constraints.

Open edX Unlocked: Designing Learning Platforms for Any Organization

Open edX is a flexible platform that can power learning, knowledge sharing and skill development for any organization. Many non-educational institutions struggle to adapt it. How do you structure content without traditional courses, measure success without grades and engage learners? In this session, we will share real-world strategies and lessons from organizations that have used Open edX for onboarding, compliance, professional development and internal knowledge management. Attendees will gain practical ideas and inspiration for using Open edX to drive engagement, impact and value across any industry.

Getting Buy-In: Strategies to Fund and Launch Open edX Initiatives Across Organizations

Many teams have great ideas for improving their Open edX platforms, from new features and integrations to custom workflows, but struggle to get approval, resources and budget from leadership. This session will explore practical strategies for building compelling proposals, demonstrating ROI and securing group funding. Drawing on real world examples, attendees will learn how to frame initiatives in ways that resonate with decision makers, align with organizational priorities and overcome common roadblocks.

5 Likes
13
  • Managing Open edX deployments at scale: Overview of Grove, lessons learned, how the PHD stack improves on Grove design, plans for the future.
  • Automating Open edX sandboxes: Overview of how automatic sandbox creation/management that we developed for Axim works.
  • MongoDB forever: There was an effort to remove the depencency on MongoDB from Open edX a few years ago, but the work on it seems to have stalled. It would be interesting to look at what was already done, what still needs to be done, what hurdles were hit, and whether removal of MongoDB remains on the roadmap.
3 Likes
14

Mastering Asynchronous Collaboration in Distributed Open edX Teams

Learn how OpenCraft, a fully distributed company, works asynchronously to deliver high-quality Open edX development. Discover practical strategies for information sharing, sprint planning, decision-making, and documentation that improve productivity and work-life balance in distributed teams, within a given organization or as a contributors in Open edX working groups.

The Open edX Project at 12+: Community Evolution and Lessons from the Trenches

From the first external open-source contributions in 2013 to todayā€™s thriving ecosystem, explore how the Open edX project evolved from an edX initiative into a mature community-driven platform, through the eyes of one of its first contributors. Learn about governance structures, technical pivots, what makes an open source project successful on the long term, and what the future of the Open edX project might be.

Building Sustainable Open edX Businesses: OpenCraftā€™s 10-Year Playbook

OpenCraft has sustained continuous contributions to Open edX for over a decade while building a sustainable business. Explore the strategies, principles, and practices that enable long-term sustainability and virtuous community collaboration in open-source education technology.

3 Likes
15

Design Tokens: Plugin Slots but for Your CSS
Design tokens donā€™t just enable deep customization; they also enable better styling for frontend plugin slots and open the doors for creating new hooks for styling and customization. Just like plugin slots allow hooking into the content of an MFE tokens allow hooking into the styling of MFEs.

Themed XBlocks with Paragon 23 (Lightning talk?)
Paragon 23 now supports loading the theme over CDN. This means we can now have XBlocks built using Paragon using the same CSS as the platform including the theme. This can make XBlock UIs meld in a lot better with the platform.

A Single Plugin for Frontend, Backend and Tutor
Iā€™ve been prototyping something like this but havenā€™t found enough time. The idea is that you have a single repo that has a Python backend plugin, a TypeScript frontend plugin, and Tutor plugin that will install both, including live-reloading when install in dev mode. This might need some actual budget for prototyping.

3 Likes
16

Demonstration of the new course outline and sidebar experience: While the course outline is relatively new, we have a new epic to redesign the outline to use sidebar for a lot of operations. This talk is based on the assumption that we are able to complete the epic before the conference.

Reuse course content: A content libraries demo, A talk about importing content from existing courses into libraries and reusing them across multiple courses using content libraries. This will demo new modulestore migrator functionality that allows authors to import content from courses into libraries as well as demo different ways of using library components in courses.

Tanstack DB: Overcome limitations of react query, A short talk about using Tanstack DB in parts of our stack where it can really shine like course outline management in course authoring. Currently, it is managed using redux, so instead of migrating to only react query, Tanstack DB could be used to manage it as a collection.

2 Likes
17

Note: as I mentioned, Iā€™m not planning to attend the conference this year if itā€™s in the US. But here are the talks that I would consider giving in theory, because these are the questions that I want answered.

Whatā€™s next for MOOCs?

AI and social media are having a profound impact on education - and often for the worse. Cheating is normalized, attention spans have been decimated, virtual proctoring is becoming nearly impossible, and those who earn a college degree are struggling to find employment in todayā€™s job market. Letā€™s look at how these trends are impacting demand for online courses, whatā€™s happening with completion/retention rates, how the role of online courses is evolving, and how an online learning platform like Open edX needs to adapt.

The Open edX Project at 12+: Technical Evolution and Architecture Lessons

Over time, a lot of major technical decisions have been made about the Open edX platform - supporting ā€œXBlockā€ python plugins for learning components, having a single codebase for both Studio and the LMS, switching to MFEs, deploying with Tutor, and so on. Which decisions have been wildly successful, and which havenā€™t? What can we learn going forward?
(with apologies to Xavier for blatantly copying / mirroring his talk)

Whatā€™s next for XBlocks?

A lightning talk about the problems with XBlock architecture (especially rendering) and some options for moving forward and improving it. e.g. can we change to use Web Components ? How can we use Paragon/React in XBlock UI? (basically an updated/summarized version of this but with new ideas.)

2 Likes
18

I think I missed posting about it earlier, but like a bunch of us Iā€™m not comfortable traveling to the US currently, so Iā€™d have to skip attending the conference next year.

  1. Facilitating contributions: The evolution of OSPR and product review processes. Where we came from, where we are now, and where we might want to go next.
  2. Evaluating offline readiness of the Open edX Android app: Results from a pilot test in Ethiopia. (Assumes that said test will be completed before the conference.)

Iā€™m struggling to come up with a third topic right now. If I think of one Iā€™ll add it before Monday end of day.

4 Likes
19
  1. A technical overview of the Learning Paths: this is similar to what @cassie proposed, but from the development perspective, so the target audience would be developers instead of product/educators.

  1. Migrating from Ansible-based deployment to Kubernetes: this is a resubmission of last yearā€™s ā€œMigrating from Ansible-based deployment to Groveā€ talk, but this time with the new stack instead of Grove. I donā€™t have much context yet, but Iā€™ll be working on this in the coming months, so it would be a good opportunity to make it a case study.

  2. Tips for local development with Tutor (probably a lightning talk): a collection of solutions to make life easier, like managing multiple environments with git worktrees (or jj workspaces), using custom scripts (e.g., to install packages), mounting shared directories for local development, applying branding and custom components to your MFEs, or adding custom MFEs to Tutor. There could also be a slide or two about how to migrate a repository from setup.py + requirements/ + Makefile to pyproject.toml + uv.lock + mise.toml.

2 Likes
20

Next Steps from the Content Libraries: As other members from Falcon mentioned, every year we have a talk about the progress on the Content Library context. This year, we can talk about the reuse of the library containers, the course import from libraries, and a ā€œhow toā€ for the Legacy Libraries migration.

Managing instances at scale: Following last conferenceā€™s infrastructure talk by Gabor (presented by Fox), we can share our experiences with the maintenance and scaling challenges we encountered with our instances. Discuss bottlenecks, scaling strategies and outrages related to the spikes that we experience.

Beyond Paragon - A discussion about the Design System for openEdx: A debate (inspired by what Casie and Ali did at the last conference) about what we want as a Design System for our platform. Sometimes I feel that many of the more complex components we develop on the frontend belong to Paragon (multiple card designs, card grids and lists, drag-and-drop components, data snippets). In other cases, we ended up with a lot of Bootstrap classes in flex divs because we sometimes lack the flexibility on Paragon, and it is easier to build it from scratch. Another point is that I feel we lack a Figma template for Paragon. Every time we work on a new set of features, we create designs with new Button styles, borders, colors, and experiences that usually look awesome but are not related to the current Paragon component set (or anything that we did previously).

4 Likes