Director, Applied AI Research - Compute & Systems

AMD AMD · Semiconductors · Santa Clara, CA · Engineering

Director of Applied AI Research at AMD, leading research at the intersection of AI, compute, and hardware. The role focuses on using agents, reinforcement learning, and recursive self-improvement to advance AMD's compute and hardware. Responsibilities include defining research agendas, leading a team, developing evaluation infrastructure, and ensuring adoption by engineering teams. The role requires a deep understanding of RL, AI-for-systems, and practical failure modes, with a track record of research impacting real systems.

What you'd actually do

  1. Define and own an applied AI research agenda focused on compute and hardware—kernel and compiler optimization, PPA and design-space exploration, and AI-for-systems, on the highest priority engineering projects at the company
  2. Build, lead, and mentor a high-caliber team of applied researchers and research engineers; set research direction, hiring bar, and standards for rigor and delivery
  3. Use recursive self-improvement loops to drive compounding, verifiable gains on concrete compute and hardware targets
  4. Design and partner with teams to own the evaluation, simulation, and verification infrastructure (fast verifiers, benchmarks) that makes applied results trustworthy and ready to adopt
  5. Confront the practical failure modes of RL and self-improving loops—reward hacking, evaluation gaming, reward-signal scaling—and build research and/or partner with research teams to detect and mitigate them

Skills

Required

  • PhD in Computer Science, Electrical Engineering, Machine Learning, or a related field, or equivalent research experience and demonstrated impact
  • Demonstrated technical leadership in applied AI/ML research
  • Deep, current expertise in reinforcement learning, including reward modeling, training pipelines
  • Strong grounding in AI-for-systems and compute/hardware problems—kernel or compiler optimization, design-space exploration, or hardware/software co-design
  • Experience building the evaluation, simulation, or verification infrastructure that applied research and training loops depend on
  • Experience leading research teams and setting direction across multiple concurrent workstreams
  • Strong cross-functional collaboration skills, with a track record of landing research inside engineering organizations
  • Track record of attracting, hiring, and developing top applied AI research talent

Nice to have

  • Ability to remain hands-on with training pipelines, kernels, or research prototyping while leading a team

What the JD emphasized

  • recursive self-improvement
  • agents
  • reinforcement learning
  • evaluation, simulation, and verification infrastructure
  • practical failure modes of RL
  • reward hacking
  • evaluation gaming
  • reward-signal scaling
  • partner deeply across silicon, architecture, compiler, software, and AI-for-hardware engineering teams
  • track record of research that reached real systems or production impact
  • practical failure modes of RL at scale

Other signals

  • leading applied research
  • recursive self-improvement
  • agents
  • reinforcement learning
  • compute and hardware optimization
  • evaluation infrastructure