Data & AI London Meetup #17

The 17th edition of Data & AI London, a 6,500-member community meetup, with three talks and Q&A. Javier Ramirez (QuestDB) - When Your Wire Protocol Becomes the Bottleneck: why, when a database stores data in columns and the client wants columns back, everything on the wire is still rows and text. Designing a new wire protocol from scratch - binary, columnar and bidirectional - and the trade-offs that came with it. Andrey Lukyanenko (Meta) - AI-driven participation in Kaggle competitions: real competitions where agents wrote code, ran long experiments and got strong results, while also making false assumptions and optimising the wrong signals. How multi-session memory, reusable skills and validation checks turn capable agents into a reliable research workflow. Alexy Golev - When the Bug Looks Like Success: Reliability Patterns for Multi-Agent AI Systems. A distributed-systems lens on agent architectures, and why an agent producing plausible, confident output is not the same as an agent working correctly. Getting in: the event space is inside the 100 Liverpool Street offices, opposite Liverpool Street mainline and Tube stations. Once in the building, the space is the Story Club, marked C on the wall; volunteers and the reception desk can direct you. Doors open 17:45, arrive by 18:00, talks start around 18:30. Pizza and refreshments on arrival. This event is currently on a waitlist on Meetup - RSVP to join it.