PGConf.dev 2026: Our team’s sessions, working groups, and key takeaways

Last week, we attended the annual PGConf.dev as a Gold-level sponsor. While most PostgreSQL conferences usually attract users and DBAs, this event draws a strong mix of contributors and community members alike, making it a unique opportunity to get proposals and patches reviewed and to connect across the broader Postgres ecosystem. 

Our team played a major role behind the scenes. Robert Haas helped organize the event, tackling the impressive feat of shaping Tuesday's content across six tracks. Additionally, Jacob Champion served on the Talk Selection Committee, and Phil Alger, Manni Wood, Álvaro Herrera, Andrew Dunstan, Floor Drees, and Euler Taveira all volunteered during the conference. 

I’ve summarized our involvement in the conference below, supplemented with comments from my team members.

Committees and working groups

Beyond the main stage, our team was embedded across working groups, panels, and community discussions throughout the week, covering everything from contributor onboarding to code of conduct policy.

Community & contributor onboarding

To kick things off, Peter Geoghegan (AWS) and I hosted a welcome breakfast for PostgreSQL community and PGConf.dev newcomers. Alongside the students and first-timers, a bunch of familiar faces dropped by to help make introductions and break the ice. Seeing everyone connect like that was incredibly satisfying!

After Robert Haas and Claire Giordano delivered the day's opening remarks, I led a panel discussion on the “human architecture of PostgreSQL” with Hari Kiran (OpenSourceDB), Jimmy Angelakos (pgEdge), Miaolai Zhou (AWS), and Valeria Kaplan (Data Egret).

The session was recorded, but the main gist was addressing volunteer burnout. We noted that a small pool of people often take on too many responsibilities, go unnoticed, and no one else steps up. We proposed ways to better recognize these volunteers as team members, acknowledged where our documentation (aka the PostgreSQL wiki) falls short, and encouraged attendees with a single, practical action they could start doing after the conference.

Following up, Robert Haas (in his role running the PostgreSQL Hacking Discord and workshops) joined Cornelia Biacsics (Microsoft), and Hari Kiran to talk more practically about “Onboarding new community members to PostgreSQL”. From this session’s notes it’s clear that while the community is experiencing significant growth, it faces challenges with discoverability, volunteer onboarding, and defining clear pathways for new contributors to get involved. To improve, the community needs to expand its focus beyond core database topics, foster local and inclusive user groups, and provide better mentorship and structural guidance for those looking to start new initiatives.

Extension Ecosystem

All but one of the sessions around extension development that Alastair Turner, David E. Wheeler, Yurii Rashkovskii, and I proposed got accepted, which means that we basically had an ecosystem track going on. Jacob Champion joined us for the session on Extending Authorisation and Authentication, all notes are recorded on the wiki.

Working Groups

Other sessions the team was involved in include the breakout on software packaging, led by Christoph Berg (Cybertec) and Devrim Gündüz. In his lightning talk on Thursday, Devrim alerted extension developers to the delivery dates of PostgreSQL 19, and asked/urged them to please update their software accordingly, for the sake of himself and Christoph as packagers. Our committers attended the closed Committers Meeting, run by Peter Eisentraut.

In the “main track” (or: “infotainment track”) Bruce Momjian delivered his “What's Missing in Postgres?” talk which can always count on plenty of questions and reactions. Robert Haas moderated a “Real-Time Patch Idea Evaluation” discussion with panelists Andres Freund, Heikki Linnakangas, and Tom Lane. The panel gave opinions on a variety of patch ideas, ranging from unsigned integer data types to bringing various extensions in-core to a shared plan cache, helping attendees to understand how experienced committers evaluate patch ideas and what the true obstacles are.

Peter Eisentraut ran the Translators and Translation Tooling meeting, where the focus was on improving the user experience for non-technical translators. Andreas Scherbaum joined Cornelia Biacsics to share ideas around (sustainable) Meetup planning and efforts. Notes from the meeting can be found on the wiki.

Jacob Champion shared in the OAuth Working Group the state of OAuth client authentication in PostgreSQL, and future work. Jacob shares: “we came away with two clear goals (token caching and bouncer support for OAuth) for the PG20 release.” Notes are on the wiki.

Peter Eisentraut attended the PostgreSQL Community Association AGM and Q&A session. The PostgreSQL Community Association is a non-profit that serves to protect the PostgreSQL project's assets such as domains and trademarks, and it usually holds its Annual General Meeting at the yearly developer conference in Canada.

In “Decentralizing safety: a proposal for local Code of Conduct response” Stacey Haysler and I presented what we think should be table stakes when it comes to Code of Conduct response handling at events. Next step: presenting it to the core team.

Conference Talks

May 21 marked the first classic conference day for PGConf.dev. Álvaro Herrera presented the REPACK work he submitted for PostgreSQL 19 in “Table repacking, done right”. Robert Haas presented his pg_plan_advice patch, and judging by the positive feedback, we can pretty safely assume it’ll make the cut for PostgreSQL 19.

Amul Sul presented “Internal data value representation in PostgreSQL”. He gave a similar presentation for our Developer U participants, the internal contributor onboarding program Andrew Dunstan talked about at PGConf.dev as well (slides). Several Developer U participants were able to join us at PGConf.dev, a great place to make connections with the people reviewing and committing their (future) patches.

With Jimmy Angelakos, our Andreas Scherbaum compiled PostgreSQL Commitfest Metrics, and presented a quantitative analysis at the event. Incorporating the feedback from the audience, they’re planning to publish a blog post with the data shortly.

Richard Guo explained how semi-joins work in PostgreSQL, and Alexandra Wang talked about “Extending Extended Statistics to Joins”. During the unconference on Friday she facilitated the discussion on Feedback-Based Query Optimization/Execution. Thanks to Corey Huinker (Apple) who posted notes here.

Vibhor Kumar presented his ORBIT framework during his session, “Postgres as an Execution Environment for AI Workflows: Failure Modes, Hooks, and Patchable Primitives” (with a deep-dive blog post promised soon!)

Lightning talks & special sessions

Different from the talks, Peter Eisentraut led a hands-on workshop to learn how to read and interpret the SQL standard and how to contribute to it, and I co-hosted a tech-focused, Toastmasters-style impromptu speaking session alongside Alastair Turner. Titled “CREATE TABLE topics ();”, the workshop challenged participants to select a technical category and deliver a 1-to-2-minute lightning talk on a surprise prompt. For example, a participant choosing "AI" might be tasked with defending or debating the prompt: "AI is flooding committers, but it can save them as well."

For the 30 Years of PostgreSQL Retrospective, Bruce Momjian and Vadim Mikheev joined fellow panelists Jan Wieck, Jolly Chen, Thomas Lockhart, Tom Lane, and hosts Jonathan Katz and Melanie Plageman, to look back at three decades of PostgreSQL: from its early beginnings, to the open source success story it is today. Other elements of celebration included the cake cutting (by Bruce) and group pictures during the conference, and the commemorative poster that was handed out and now proudly hangs in my home office.

Unconference

The Friday was reserved for an “unconference”. An unconference is participant-driven, the agenda is created by the attendees at the start of the day rather than by the organizers.

Microsoft’s Nazir Bilal Yavuz suggested “Postgres CI” as a topic for an unconference session. Peter Eisentraut joined the session, urged by the fact that the existing CI solution Cirrus CI will shut down on June 1st. The participants agreed to migrate to GitHub Actions and will focus on getting a minimal patch in as soon as possible, with the remaining features getting filled in later. Another example of how real issues get solved at PGConf.dev.

The whole week there was hallway chatter around the increase of AI (assisted) patches and reviews, and how the committer team is going to deal with that with their real human constraints. Joe Conway (AWS) suggested an unconference session called “Code, AI, and you”, and the main room was filled.

What's next?

We look forward to seeing what patches will actually ship with PostgreSQL 19. No doubt also attendees left the conference inspired with ideas for the next 20 release. PGConf.dev already announced their 2027 dates, May 11-14, in Montreal, and EDB will be there for sure.