Software Engineer, Core AI Compiler & Runtime

You will design and build the compiler and runtime that power one of the most advanced and widely deployed AI platforms on Earth—running neural networks in millions of Tesla vehicles and Optimus robots. Full control of the hardware means you co-design the software stack to extract maximum performance, latency, and efficiency that generic frameworks cannot reach. Production models directly drive priorities: every optimization you ship immediately enables more capable autonomy, more complex robot behaviors, and higher-throughput data-center inference. This is first-principles systems work at the metal including deleting unnecessary abstractions, inventing high performance APIs and DSLs, and closing the loop between silicon capabilities and real-world AI at an unprecedented scale.

Responsibilities

Write, debug, and maintain robust compiler and runtime software for Tesla AI inference across FSD, Optimus, and data-center workloads Give AI developers direct access to performance-critical hardware features through new APIs and domain-specific languages Design compiler and runtime features that enable efficient distributed inference on Tesla hardware Integrate the deployment stack with modern ML frameworks (PyTorch, JAX) while keeping the path to silicon as short and efficient as possible Analyze and resolve functional and performance issues on massively parallel systems Deploy and optimize AI models that run in autonomous vehicles, humanoid robots, and offline inference pipelines where results show up in the real world

Requirements

Demonstrated expertise with ML compilers and runtimes (MLIR, LLVM, XLA, PJRT, TensorRT, or equivalent) Experience developing domain-specific languages or high-performance programming models (Triton, cuTile, Pallas, or similar). Strong familiarity with CPUs, GPUs, and modern AI accelerators Deep knowledge of computer architecture, distributed systems, networking, and collectives Proficiency in modern C++ and solid systems programming skills Degree in Engineering, Computer Science, or equivalent in experience and evidence of exceptional ability