Senior Software Engineer — ML Inference & Numerics
Core
Build agentic workflows and numerical optimization tools to accelerate LLM inference on custom silicon, ensuring model quality and performance.
Role type
Senior Software Engineer (ML Inference & Numerics)
Builds
Agentic harnesses for model exploration, SDK components, and numerical validation software for generative AI inference.
Domain
Generative AI inference, custom silicon, hardware acceleration, numerical optimization.
Deliverable
production ML models
Required skills
LLM inference architecture, Python, agentic tooling, numerical optimization, model quantization, distributed inference frameworks, hardware-software co-design, system design, testing, API design.
Preferred skills
ML compilers, accelerator SDKs (TPU, Trainium/Inferentia, Tenstorrent, Groq, Cerebras), model compression techniques, contributions to vLLM or SGLang.
Responsibilities
Build agentic workflows to inspect models and run experiments; optimize precision, accumulation, and quantization choices; create conversion and emulation software for numerical validation; own design, implementation, and delivery of SDK components; collaborate with hardware architects on numerical format impacts.
Seniority
Senior, hands-on IC
