Director, AI Research - Recursive Self-Improvement

Lead AMD’s recursive self-improvement (RSI) agenda using reinforcement learning to measurably improve hardware design, kernel and compiler optimization, and engineering workflows. Build and mentor a research team, set technical direction, and drive RL research on internal targets like design-space exploration and code/workflow generation. Own verification and evaluation infrastructure, tackle RL failure modes (reward hacking, evaluation gaming), and partner cross-functionally to embed results into real engineering flows.