ML Systems Engineer – Large-Scale Training & Inference

About the role

Periodic Labs is building AI‑driven scientific models that accelerate breakthroughs in materials, energy and beyond. As an ML Systems Engineer you will design and own the infrastructure that powers large‑scale training, inference and reinforcement‑learning workloads for cutting‑edge AI agents.

Key responsibilities

• Build and optimise large‑scale training and reinforcement‑learning pipelines, ensuring correctness and high throughput.
• Develop high‑performance inference and serving systems for AI models.
• Design distributed runtimes and scheduling frameworks for complex ML workloads.
• Create secure sandboxing and execution environments for safe model execution.
• Optimise memory usage, GPU kernels and inter‑GPU communication for maximum efficiency.
• Improve scalability, reliability and overall performance across the ML systems stack.

Required profile

• Strong systems‑programming and performance‑engineering background.
• Proven experience building high‑performance ML infrastructure at scale.
• Ability to own complex technical problems end‑to‑end.
• Solid coding judgement and ability to work with AI agents on design, implementation, testing and debugging.
• Bachelor’s degree or equivalent practical experience.
• High ownership, fast execution and passion for advancing AI systems.

Required skills

• Megatron‑LM
• TorchTitan, FSDP, veRL, Slime
• Ray (and optionally Monarch)
• SGLang (and optionally vLLM, TensorRT‑LLM)
• Secure execution environments, containers, virtualization, code sandboxing
• CUDA, Triton, CUTLASS, CuTe
• NCCL, NVLink, InfiniBand, RDMA, GPUDirect RDMA

What we offer

• Base salary $250,000–$350,000 per year plus equity.
• Visa sponsorship for eligible candidates.
• On‑site work in Menlo Park, CA (future San Francisco office).