Summary
Julia Kiseleva is an AI scientist-founder with 15+ years translating cutting-edge research into deployable products, currently founding a stealth startup after serving as Chief Scientist at Please.ai and a long tenure at Microsoft Research. She specializes in dialog systems, human-AI feedback loops, and multi-modal agentic interaction, with a track record of moving models from evaluation into real-world, user-aligned systems. An organizer of NeurIPS competitions and advisor to startups, she combines rigorous academic training (PhD, Eindhoven) with hands-on product and startup experience dating back to co-founding UserSat. Based in San Francisco, she’s as comfortable designing experiments and benchmarks as she is building teams and company strategy, and she often focuses on trustworthy, grounded AI—an emphasis that informs both her research and product decisions.
8 years of coding experience
18 years of employment as a software developer
MS, Applied Mathematics, MS, Applied Mathematics at Saint Petersburg State University
Doctor of Philosophy (Ph.D.), Doctor of Philosophy (Ph.D.) at Eindhoven University of Technology
RUSSIR 2008
English, Russian, Dutch, Spanish