Director, AI Research - Recursive Self-Improvement

What you'll do

• Define and own the research agenda for recursive self-improvement in hardware design, kernels, compilers, and engineering workflows.
• Build, lead, and mentor a high-caliber team of AI researchers and engineers while setting technical standards.
• Drive reinforcement learning research to achieve measurable improvements in kernel optimization and design-space exploration.
• Design and manage verification infrastructure, including simulators and benchmarks, to ensure safe and reliable self-improvement loops.
• Develop programs to detect and mitigate common RL failure modes like reward hacking and evaluation gaming.
• Partner with silicon, architecture, and compiler teams to integrate AI research into internal development flows.
• Track measurable goals such as verified performance gains and cycle-time reductions for the RSI program.
• Communicate progress, risks, and roadmap implications of the research agenda to executive leadership.

What we're looking for

• PhD in Computer Science, Electrical Engineering, Machine Learning, or a related field, or equivalent research experience and demonstrated impact.
• Demonstrated technical leadership in AI/ML research with a track record of published work or production systems.
• Deep expertise in reinforcement learning, including reward modeling, training pipelines, and scaling failure modes.
• Hands-on understanding of AI-for-systems problems like code generation, compiler optimization, and hardware/software co-design.
• Experience building coding RL loops.
• Proven experience leading research teams and setting technical direction across multiple concurrent programs.
• Ability to remain hands-on with training pipelines, kernels, or research prototyping while leading a team.
• Strong cross-functional collaboration skills to embed research into engineering organizations.