Software Engineer, Reinforcement Learning, AI Infrastructure

As a Software Engineer on Tesla AI’s RL Infrastructure team, you will build the systems that make large-scale reinforcement learning fast, stable, and flexible for ML engineers working on real-world AI problems. Your work on RL training will be critical for scaling Robotaxi and Optimus. We're looking for someone with practical experience in GPU inference performance, RL training infrastructure, or both. Your work will directly impact iteration speed, training throughput, system reliability, and the ability of ML engineers to move from prototype to large-scale experiments quickly.

Responsibilities

Build infrastructure for large-scale reinforcement learning Improve performance and scalability of RL workloads Optimize training and inference latency, throughput, memory usage, and hardware utilization Develop tools for benchmarking, profiling, debugging, and monitoring Work closely with ML engineers to productionize new methods and workflows Debug bottlenecks across kernels, runtimes, networking, and distributed systems Improve system reliability, correctness, and engineering quality across the RL stack

Requirements

Industry experience in GPU inference, writing GPU kernels, and/or RL training infrastructure Practical programming experience in Python and/or C/C++ Experience working with ML training frameworks (ideally PyTorch) Strong understanding of performance engineering and distributed systems Experience with PyTorch, CUDA, NCCL, Triton, Ray, or similar tools Experience profiling and optimizing CPU-GPU interactions (pipelining computation with data transfers, etc.) Proficient in system-level software, in particular hardware-software interactions and resource utilization Ability to solve difficult systems problems with minimal oversight