About the position
As a Software Engineer within our Autonomy teams, you will contribute to one of the most advanced and widely deployed AI Platforms in the world for Autopilot and our Humanoid Robot, Optimus. In this role, you will be responsible for the internal working of the AI inference stack and compiler running neural networks in millions of Tesla vehicles and Optimus. You will collaborate closely with the AI Engineers and Hardware Engineers to understand the full inference stack and design the compiler to extract the maximum performance out of our hardware. The inference stack development is purpose-driven: deployment and analysis of production models inform the team's direction, and the team's work immediately impacts performance and the ability to deploy more and more complex models. With a cutting-edge co-designed MLIR compiler and runtime architecture, and full control of the hardware, the compiler has access to traditionally unavailable features, that can be leveraged via novel compilation approaches to generate higher performance models.
Responsibilities
- Write, debug and maintain robust software for Autopilot and Humanoid robot AI inference (Export / Compiler / Runtime) stack
- Work with AI/runtime/OS/driver teams to enable efficient compilation and execution of the AI Models
- Bring up compiler backends for future versions of Tesla HW accelerators
- Focus on advanced compiler algorithms (heterogeneous scheduling, accelerator instruction scheduling, memory placement, etc.)
- Keep up-to-date and collaborate with ML/compiler community to keep the stack compatible with latest developments
Requirements
- Proficient C/C++ programming including modern C/C++ (C++14/17/20)
- Basic Python proficiency
- Prior experience working with compilers, with a focus on middle and backend
- Experience with ML compilers and frameworks is a plus (e.g. MLIR, XLA, TensorRT, LLVM, JAX, TVM)
- Degree in Engineering, Computer Science, or equivalent in experience and evidence of exceptional ability
