ML Infrastructure Engineer

White Circle is an AI Safety company seeking an ML Infrastructure Engineer to build robust systems for LLM post-training, RL, evaluation, and inference. The role involves designing data control systems, tuning training and inference pipelines, and working closely with researchers. The company has raised $11M and processes over 100M API calls monthly.

• Build scalable RL and post-training pipelines including smoke tuning runs and approach ablations.
• Design data control systems governing training data flow.
• Tune training and inference end-to-end for high throughput across networking, memory, compute scheduling, data loading, storage, checkpointing, and I/O.
• Investigate infrastructure choices affecting learning dynamics and training stability.
• Build infrastructure for model iteration: experiments, artifacts, evals, dashboards, and cost visibility.
• Work on inference infrastructure for post-training and evaluation loops.
• Build and improve agentic development environments and multi-agent orchestration.

• Experience designing distributed RL/post-training systems at scale.
• Proficiency in PyTorch or JAX.
• Strong Python skills including concurrency, async, multiprocessing, and performance optimization.
• Ability to debug distributed GPU workloads across CUDA, NCCL, networking, storage, and scheduling.