Senior Software Engineer, Inference Performance
Core
Build the software stack (abstractions, kernels, libraries) that takes AI models from code to hardware to enable peak performance on custom silicon.
Role type
Senior IC software engineer (inference performance)
Builds
High-performance generative AI inference systems for hyperscaler and neocloud data center customers
Domain
AI/ML inference acceleration, custom silicon, hardware-software co-design
Deliverable
production ML models
Required skills
hardware performance optimization, Python, Rust, C++, C, LLM inference end-to-end understanding, agentic development, API/library design, profiling and benchmarking
Preferred skills
embedded/bare-metal programming, MLIR/compiler frameworks, kernel DSLs (Triton, Pallas, NKI, CuTe), LLM serving stack contributions (vLLM, SGLang)
Responsibilities
Design APIs and kernel abstractions in Python SDK or Rust; optimize LLM and multimodal models for peak performance on real silicon; profile and fix bottlenecks across kernels, compiler, and runtime; build agentic performance loops with profilers and traces; define performance measurement standards; design abstractions for future hardware generations
Seniority
Senior, hands-on IC