Andy Pavlo Leaves CMU Full-Time to Lead New ClickHouse Labs Research Group

Carnegie Mellon database professor Andy Pavlo is joining ClickHouse to found and lead ClickHouse Labs, an industry research team modeled on the likes of IBM Research and Microsoft Research. Pavlo, a longtime tracker of DBMS internals, notes he has followed ClickHouse since its 2016 open-source debut, when its C++ codebase and SIMD-based vectorized execution set it apart from the JVM-based analytical engines that dominated open source at the time. Rather than operating as a detached lab, the group is meant to work directly with ClickHouse’s engineers, customers, and partners to move research into production.

A near-term priority is clearing the backlog of optimizations the engineering team has prototyped but never fully validated or shipped, then building on that to push the system further. The team will also support ClickHouse’s managed PostgreSQL offering, giving the group a foothold in both transactional and analytical database problems.

The more open-ended research direction is the intersection of databases and AI agents, framed as two questions: how a DBMS should be designed to better serve agents, and how agents can automate the building of database systems themselves. Pavlo argues that ClickHouse’s relational foundation leaves it well positioned for these data-intensive workloads, with hardware, algorithms, data structures, and execution strategies all in scope.