Reflections on OpenSearchCon NA 2026

OpenSearchCon North America 2026 was in San Jose, and I’ve been meaning to write something about it ever since I got home, but between catching up on client work and following up with the people I met there it has taken me a couple of weeks to actually sit down and do it.

My talk

I gave a talk called Leveraging LDP for High-Trust OpenSearch UBI, which was about the awkward fact that the behavioral data that makes search better is also the data your users are most sensitive about, and how local differential privacy lets you collect that data without ever holding anyone’s true behavior. The idea is that each client randomizes its own User Behavior Insights events before they leave the device, so any single event is deniable while the noise mostly cancels out once you aggregate enough of them, and the signals you need for relevance work are still there when you go looking for them.

The conversations I had after the session turned out to be my favorite part of giving the talk. Many of the teams I talked with are already collecting behavioral data from their search applications and are starting to feel uneasy about holding onto it, which is exactly the problem I was hoping the talk would help with. If you missed the talk, or saw it and want to try things yourself, the code and the notebook that walks through applying LDP to UBI data end to end are on GitHub at jzonthemtn/ldp-for-search.

The people

The sessions were good, but what I’ll remember most is how many people I got to meet over those three days, some of whom I had only ever known as GitHub handles and many of whom I had never crossed paths with before. I talked with fellow maintainers and with teams who are still working out where OpenSearch fits for them, and I learned something about how the project gets used in the real world from nearly every one of those conversations.

Migrations came up a lot, which I was happy about since I’m a maintainer on opensearch-migrations, and it was great to see a full room for the Migration Assistant session and to hear Aiven talk about making it portable across clouds. Moving a cluster without quietly breaking relevance along the way is something I spend a lot of time on with clients, so it’s encouraging to see the tooling for it getting this much attention from more than one company. I also sat in on a session where Eric Pugh of OpenSource Connections talked about how a technical advisory group for search might fit between the TSC and the maintainers, offering guidance and feedback on things like RFCs while leaving the actual decisions with the people who own the code.

Being an ambassador and on the TSC

This was my first OpenSearchCon since being named an OpenSearch Ambassador and elected to the Technical Steering Committee, which I wrote about in an earlier post, and having those roles changed how I spent my time there, since I found myself paying less attention to my own talk and more attention to the problems people kept bringing up, because those are the things worth carrying back to the project.

I’m really excited about both roles. The ambassador side is about helping people find and adopt OpenSearch, and the conference made it obvious how much interest there is right now, while the TSC side is about technical direction, where I want to keep pushing on privacy as something the whole project thinks about rather than something each plugin handles on its own. Talking with users in San Jose reinforced my sense that privacy questions show up in production well before they show up in feature discussions, and I’d like the project to be ahead of that rather than reacting to it.

Thanks

Thanks to the organizers, the other speakers, and everyone who came to my talk or stopped me in a hallway to say hello, because you all made it a really enjoyable few days, and I’m already looking forward to the next OpenSearchCon.

If you’re working through privacy in your search data, planning a migration to OpenSearch, or just want to pick up a conversation we started in San Jose, please reach out.