MTS - Agent Research
Core
Set research direction and spearhead development of self-improving agents to complete difficult, long-running engineering tasks for hardware infrastructure.
Role type
Senior IC research lead (agentic systems)
Builds
Self-improving agent systems, custom models, and evaluation frameworks for engineering workflows
Domain
Frontier AI infrastructure, hardware co-design, inference systems
Deliverable
production ML models | product features
Required skills
research judgment, engineering ability, agent development for complex tasks, LLM post-training (SFT/RL), experiment design under constraints, technical direction
Preferred skills
ASIC design, accelerator kernels, large-scale compute infrastructure
Technologies
SFT, RL, distillation, in-context learning, agent harnesses
Responsibilities
Own research agenda for agentic systems, work with domain engineers to diagnose agent failures, develop systems for long-running tasks (memory, context, tool use), build evaluations for correctness and efficiency, train custom models using SFT/RL, scale concurrent agent experiments
Seniority
Senior, hands-on IC