ML Infra Engineer: Scale RL Pipelines & Inference

Moonfire; Paris (TX); On-site; USD 115,000 - 173,000; Full time

Benefits offered by this job

Equity

Flexible time off

Relocation package

Medical insurance

Learning & development

Equipment & tools

Team off-sites

Job summary

White Circle is seeking an ML Infrastructure Engineer to build the systems behind LLM post-training, RL, evaluation, and inference workflows. You’ll work closely with researchers on GPUs, training loops, data control systems, and the infrastructure decisions affecting model learning and product quality.

You will design and tune end-to-end training and inference stacks, build agentic development environments, and ensure scalable, reproducible experiments.

Qualifications

• Designed, built, or maintained distributed RL/post-training systems at scale.
• Fluent in PyTorch or JAX.
• Proficient in Python, including concurrency and performance optimization.
• Experience debugging distributed GPU workloads (CUDA, containers, NCCL, networking).
• Experience with profiling tools (py-spy, PyTorch profiler, Nsight, perf).
• Familiarity with inference stacks (vLLM, SGLang, TensorRT-LLM, Dynamo).
• Ability to reason how infrastructure affects learning and eval quality.

Responsibilities

• Build robust, scalable RL and post-training pipelines for quality testing and ablations.
• Design data control systems governing data visibility and training flows.
• Tune training and inference end-to-end for high throughput (networking, memory, I/O).
• Investigate how infra choices affect learning dynamics and evaluation quality.
• Build infrastructure for model iteration: experiments, artifacts, evals, dashboards, reproducibility.
• Develop inference infrastructure supporting post-training and evaluation loops.
• Create agentic development environments: coding-agent harnesses, tool integrations, sandboxes.
• Collaborate with team to plan steps, discuss tradeoffs, and maintain communication.

Skills

Python programming

Distributed systems

Performance optimization

Research collaboration

Tools

PyTorch

JAX

CUDA

Kubernetes

Slurm

Ray

NCCL

UCX

Job description

White Circle is seeking an ML Infrastructure Engineer to build the systems behind LLM post-training, RL, evaluation, and inference workflows. You’ll work closely with researchers on GPUs, training loops, data control systems, and the infrastructure decisions affecting model learning and product quality.

You will design and tune end-to-end training and inference stacks, build agentic development environments, and ensure scalable, reproducible experiments.