Lead design, implementation, and operation of training performance optimizations for VLM, VLA, and foundation-model training.
Own the technical strategy and multi-quarter roadmap for reinforcement-learning frameworks and performance optimization.
Profile and diagnose throughput bottlenecks across compute, compiler, framework, and cluster-networking layers.
Partner with ML researchers, data engineers, and hardware engineers on requirements and architectural decisions for large-scale training infrastructure.
Build custom kernels and benchmarks to measure and validate performance improvements.
Mentor engineers and raise the team’s technical bar through design reviews and best practices.
Requirements
7+ years of total experience, including 4+ years working on large-scale model training platforms and leading projects end to end.
Deep experience with distributed training frameworks such as Megatron-LM and TorchTitan, and efficient use of GPU or accelerator clusters at scale.
Experience with RL frameworks such as Verl and MILES and algorithms including GRPO and PPO.
Experience with collective communication libraries such as NCCL and optimizing communication/compute overlap in multi-node training.
Strong command of profiling tools including NVIDIA Nsight, PyTorch Profiler, or XLA/JAX profiling.
Proficiency in Python or C++.
Experience with ML compilers, performance tooling, or hardware/software co-design involving GPU, TPU, or accelerator architectures.
Ability to influence across teams and communicate technical tradeoffs to researchers and engineers.
Salary: $254k - $350k/yr
Zoox
Zoox is transforming mobility-as-a-service by developing a fully autonomous, purpose-built fleet designed for AI to drive and humans to enjoy.