Software Engineer, LLVM Compiler - MTIA

Job Description

Own a key GenAI inference framework area on MTIA, designing and shipping production features across serving, distributed inference, graph execution, compilation, and PyTorch integration. Lead cross-team programs and optimize performance, accuracy, stability, and test coverage while partnering with compiler, kernel, and silicon teams.

Requirements: Requires at least five years of hands-on experience with GenAI frameworks, training or inference, or production model-serving systems, plus proficiency in Python, C++, or Rust and experience optimizing inference and runtime execution. Candidates should have led cross-team infrastructure delivery and be able to debug and maintain numerical accuracy and stability; experience with accelerator platforms, distributed serving, and low-precision numerics is preferred.

Key Skills: Generative AI Inference, Python, C++, Rust, LLM Serving, Distributed Inference, PyTorch, LLVM, Compiler Optimization, Graph-Mode Execution, KV-Cache Management, Performance Profiling, Numerical Accuracy, Quantization, CI and Test Strategy, Technical Leadership

Benefits: Bonus, Equity, Benefits