Andy Pavlo joins ClickHouse to launch ClickHouse Labs

Andy Pavlo, who has taught computer science at Carnegie Mellon University since 2013 and built his career studying the internals of database management systems, announced he is joining ClickHouse to establish and lead a new research team called ClickHouse Labs. Pavlo says he has tracked ClickHouse since its 2016 open-source debut, when he initially assumed it was vaporware because it already used C++ and SIMD-based vectorized query execution at a time when most open-source analytical databases were JVM-based and lacked those optimizations. He has followed the project ever since, including wearing a ClickHouse shirt in his first remote database lectures in 2020, and says the company's co-founders invited him to start the research group.
ClickHouse Labs will not work as an isolated unit that hands ideas to engineering from the outside; instead it will collaborate directly with ClickHouse engineers, customers, and industry partners. The team will also work with ClickHouse's PostgreSQL group to help its managed PostgreSQL service compete on performance and reliability, giving the lab a base spanning both transactional and analytical database work. Pavlo frames the goal as building an industry research organization on the model of IBM Research and Microsoft Research: producing work with scientific value, turning the strongest ideas into shipped technology, and training database researchers.
An immediate priority is clearing a backlog of optimizations ClickHouse's engineering team has already explored but has not had time to fully validate and put into production; Pavlo wants to use that as a launching point for further research. He also points to the company's existing publishing record, including a 2024 VLDB paper on ClickHouse's core architecture that he assigns as required reading to his students, as a foundation to build on. Longer term, the lab plans to investigate how databases like ClickHouse and PostgreSQL should evolve alongside AI agents, looking at it from two directions: what a DBMS needs to look like to better support agents, and how agents themselves might help automate and improve DBMS development. Pavlo says he has no answers yet since this is research, but is confident ClickHouse's relational model foundation leaves it well positioned for that shift. The announcement does not give a start date for Pavlo at ClickHouse, does not say whether he is leaving or keeping his Carnegie Mellon post, and discloses no team size, budget, or compensation details for the new lab.
Key facts
- Andy Pavlo, a Carnegie Mellon computer science professor since 2013, is joining ClickHouse to found and lead a new research team called ClickHouse Labs.
- The ClickHouse co-founders invited Pavlo to set up the group, which will work directly with ClickHouse's engineers, customers, and industry partners rather than operate in isolation.
- Labs will also work with ClickHouse's PostgreSQL team to help its managed PostgreSQL service compete on performance and reliability.
- An early priority is validating and productizing a backlog of optimizations ClickHouse engineers have already explored but not shipped.
- The group plans to research how databases should support AI agents, and how agents could help build and improve database systems, though Pavlo says he has no answers yet.
Why it matters
A well-known academic database researcher moving in-house to lead a formal research arm signals ClickHouse is betting on sustained, published research rather than ad hoc engineering as a way to stay ahead technically, following the model of corporate labs like IBM Research and Microsoft Research.
Who it affects
ClickHouse users and its PostgreSQL managed-service customers, since Labs will work on both the core analytical DBMS and the PostgreSQL offering; also the wider database research community, given Pavlo's academic ties and his plan to keep training researchers.
How to use it
There is no product to use yet. ClickHouse Labs is a new internal research organization, not a released feature or service; the announcement lists no pricing, availability, or release timeline.
How solid is it
The announcement comes directly from Pavlo on ClickHouse's own blog, so the fact of the hire and the group's stated mandate are well sourced, but it is a first-person mission statement rather than an outcome: there are no results, roadmap dates, team size, or budget disclosed yet.
Risks and caveats
Key operational details are missing: no start date for Pavlo at ClickHouse, no statement on whether he keeps his Carnegie Mellon professorship, and no information on Labs' headcount, funding, or near-term deliverables. The AI-agent research direction is explicitly described as open questions rather than planned features.
“It will not operate as an isolated research organization that throws ideas over the wall to engineering.”
— Andy Pavlo