Software Engineer, ML Training Performance (AI)

Location: Foster City, California, United States; on-site

Company

Zoox develops autonomous vehicle technology and large-scale machine learning infrastructure.

What you will do

• Design, implement, and operate ML training performance optimizations for VLM, VLA, and foundational models.
• Develop distributed training approaches including data, tensor, and pipeline parallelism, mixed precision, gradient checkpointing, and GPU kernels.
• Optimize reinforcement learning frameworks and performance for driving foundation models.
• Profile and debug throughput bottlenecks across compute, compiler, framework, and cluster-networking layers.
• Collaborate with ML researchers, data engineers, and hardware engineers on large-scale training infrastructure architecture.
• Implement custom kernels and build benchmarks to validate training-performance improvements.

Requirements

• 4+ years of total experience, including 2+ years working on large-scale model training platforms.
• Experience with distributed training frameworks such as Megatron-LM or TorchTitan and efficient use of GPU or accelerator clusters at scale.
• Experience with reinforcement learning frameworks such as Verl or MILES and algorithms including GRPO or PPO.
• Experience with collective communication libraries such as NCCL and multi-node communication/compute overlap optimization.
• Experience with NVIDIA Nsight, PyTorch Profiler, or XLA/JAX profiling tools.
• Proficiency in Python or C++, plus experience with ML compilers, performance tooling, or hardware/software co-design.

Culture & Benefits

• Full-time employment in an on-site software systems department.
• Paid sick leave, vacation, and bereavement leave, plus unpaid time off.
• Health, long-term care, disability, and life insurance.
• Compensation may include Amazon RSUs, Zoox Stock Appreciation Rights, and a possible sign-on bonus.