Skip to content

· Synx Data Labs

Community Over Code Asia 2026 Beijing: A Recap

Our CTO recaps Community Over Code Asia 2026 in Beijing: conversations with ASF leaders, Cloudberry's path to top-level project, and the students.

At Synx Data

I’m just back from Community Over Code Asia 2026 in Beijing (August 7-9), and I wanted to share a few highlights while they’re still fresh — following up on our pre-event announcement about the talk I gave there.

Shine Zhang presenting at Community Over Code Asia 2026

Our talk

I presented “Build Once, Run on Any Linux: SynxDB CE for Apache Cloudberry (Incubating)” — a candid walkthrough of the engineering behind SynxDB Community Edition’s portability strategy: bundling our own toolchain, targeting glibc 2.17 as the ABI floor, statically linking the C++ runtime while keeping glibc dynamic, and verifying portability as a CI gate. The result is one binary tree that produces both RPM and DEB artifacts and runs on any Linux with a recent-enough glibc.

Slide from Shine's talk: statically linking the C++ runtime

Apache Cloudberry’s graduation is underway

The biggest community news out of this conference, for us: Apache Cloudberry is now in the process of graduating from the Apache Incubator to a full Apache Top-Level Project. Several of us at Synx Data Labs were among the original initiators of Cloudberry, so watching it move through this milestone — through the Incubator’s governance process, with a self-sustaining, community-driven contributor base — is genuinely gratifying, even with the process still underway. It’s the kind of validation that only comes from doing the unglamorous work of building an open community around a codebase, not just shipping the code.

Conversations that stuck with me

Community Over Code Asia is as much about the hallway conversations as the talks, and this year was no exception.

I spent a good chunk of time with Justin McLean — ASF Board Member (since 2020), VP of the Apache Incubator (since 2018), and VP Legal — but not talking about Cloudberry at all. We ended up deep in a conversation about MCP and Claude Skills: I pushed him on why he’d built an MCP server rather than starting with Skills, which turned into a genuinely interesting back-and-forth about the tradeoffs. I also asked to fork his repo on the spot — turns out I was the first person to fork it.

I also spent time with Jerry Tan (Tan Zhongyi), an independent consultant who’s chaired the AI-Infra track at Community Over Code Asia for six consecutive years and serves as Deputy Secretary-General of the China Open Source Promotion Alliance (COPU). Jerry’s been watching the open-source market in Africa closely and had a lot to share on that front. He also made an observation I’ve been chewing on since: booth traffic at conferences like this one hasn’t been as strong as in years past, which he chalks up to the broader industry’s attention shifting toward AI.

And good conversations with Dianjin Wang — from Kuke, and a collaborator in the Apache Cloudberry (Incubating) community (LinkedIn) — plus a look at some of the other Apache incubating projects making their own way through the process.

The students

Some of the most memorable conversations weren’t with industry veterans at all — they were with the students who stopped by. The English and foreign-language majors we talked with were full of energy, but candid about a real worry: whether AI will make their skills obsolete. The computer science majors were candid in a different way — a lot of them spoke openly about how tough the job market feels right now, and how hard it’s been to land that first opportunity. Neither conversation was easy to hear, but both were honest, and worth sitting with.

Group photo with ASF community members at Community Over Code Asia 2026

At the booth

We also had a steady stream of visitors at the Synx Data Labs / Cloudberry booth throughout the event, including a few conversations with potential investors interested in what we’re building. One question stuck with me: an investor asked about our approach to encryption for data solutions that need to sustain extremely high throughput — on the order of 2TB/sec. Nothing to announce yet, but it’s a good signal for where the ecosystem sees this heading.

Thanks

Thanks to the ASF and the Community Over Code Asia organizers for putting together another great event, and to everyone who stopped by our talk or our booth to chat. Beijing was a good week — see you at the next one.