Member of Technical Staff, AI-Driven Compilation
$275k - $315k • Remote • San Francisco • FullTime
Posted 1mo ago
Job Location
San Francisco
Tech Stack
Remote Work Policy
Fully remote
Employment Type
FullTime
Categories
AI Agent Engineer
About the job
SF Tensor is revolutionizing high-performance compute for AI by rebuilding the entire stack, from hardware to cloud. We are developing a Kernel Optimizer that automatically finds the fastest code for any vendor or cluster topology, and a Model Foundry to manage runs, simplify research, and enable workload portability across clouds and chips. Our mission is to eliminate the compute tax, making AI development faster, cheaper, and more accessible. We are backed by prominent investors and industry leaders, and we seek individuals who believe that advancements in AI are contingent on leaps in compute infrastructure.
Responsibilities
- Design and implement RL systems for searching vast program spaces, including instruction selection, schedules, tile sizes, fusion strategies, and phase ordering.
- Build agentic compilation loops leveraging LLMs for IR reasoning, transformation proposals, and learning from measured results.
- Design search and credit-assignment machinery based on measured latency on real hardware and formal correctness proofs.
- Create representations and embeddings of compiler IR suitable for learned optimization.
- Develop training infrastructure for compiler optimization agents, including rollout throughput and distributed evaluation on real silicon.
- Improve transfer and cold-start performance for search on unfamiliar targets and new ISAs.
- Integrate learned components into the shipping pipeline by collaborating with compiler and kernel engineers.
- Conduct rigorous experiments, analyze results, iterate, and potentially publish or open-source findings.
Requirements
- Strong background in reinforcement learning with hands-on agent training experience.
- Experience building LLM agents, tool use, and other agentic systems.
- Familiarity with GPU programming concepts and willingness to work at the ISA level.
- Proficiency in PyTorch or JAX.
- Ability to design and execute experiments yielding trustworthy results.
- Experience with ML compiler stacks (XLA, TVM, Triton, MLIR) or LLVM backends.
- Background in program synthesis, superoptimization, combinatorial search, formal methods, SMT solvers, or verified compilation.
- Familiarity with RLHF, reward modeling, or preference learning.
- Research contributions in RL or learned optimization.
- Familiarity with GPU performance optimization, profiling, and microbenchmarking.