I am a Research Scientist at Google DeepMind, where I co-founded and co-lead the Machine Learning for Computer Architecture team. My research focuses on building autonomous AI agents that write, optimize, and formally verify performance-critical software and hardware.
Much of my recent work centers on agentic code and system optimization. , our LLM-driven optimizer deployed across Google's warehouse-scale production fleet, has landed 6,400+ commits and reclaimed the equivalent of hundreds of thousands of CPU cores. I also leverage to discover novel compiler heuristics, interprocedural code layouts, and FHE kernels for TPUs. Beyond optimization, I conceived and led the multi-agent effort that built a cycle-accurate RISC-V CPU simulator↗ from scratch capable of booting a full operating system—demonstrating autonomous end-to-end engineering of complex systems.
Earlier, I co-authored , one of the earliest works demonstrating that large language models can iteratively improve their own outputs through self-feedback—a foundational step toward recursive self-improvement cited 7,000+ times.
My research has been recognized with multiple IEEE Micro Top Picks awards, induction into the ISCA Hall of Fame, and elevation to Senior Member of IEEE and ACM. I also co-founded MLArchSys↗, the flagship workshop on ML for architecture and systems co-located with ISCA since 2020. I earned my Ph.D. from Georgia Tech, supported by the Microsoft Research and Qualcomm Innovation Fellowships.