ML Infrastructure Engineer: Training & Inference, Equity

Physical Superintelligence; Boston (MA); Hybrid; USD 140,000 - 230,000; Full time

Job summary

Physical Superintelligence in Boston is hiring an ML infrastructure engineer to own and evolve the training and inference stack powering our scale-out research. You will design, implement, and operate distributed multi-GPU training jobs, model-serving pipelines, and the tooling researchers rely on to run experiments efficiently.

You will collaborate with engineering on capacity planning, observability, and reliability, and you will be the go-to person when pipelines stall or workloads fail.

Qualifications

• Three or more years building and operating ML training or inference infrastructure in production.
• Hands-on experience with distributed training (multi-GPU or multi-node) and model-serving systems.
• Strong software engineering fundamentals for building reliable services.
• ML fluency to work productively with researchers even if you don't design algorithms.

Responsibilities

• Own the training and inference infrastructure that Core AI depends on: distributed training jobs, GPU scheduling, and model-serving systems for both proprietary models and self-hosted inference.
• Build the tools and abstractions AI researchers use to launch training runs, iterate on inference providers, and route workloads across models.
• Partner with Engineering on capacity planning, observability, and reliability for GPU and inference infrastructure.
• Debug and harden the training and inference stack under real load. Egress failures, stalled retries, and routing edge cases are your problem to close.
• Stay hands-on. You write the code, not just the design doc, and you are the first call when a training job stalls or an inference path breaks.

Skills

Distributed training

PyTorch

Ray

Model-serving

Tools

vLLM

SGLang

Triton

CUDA

Job description

Physical Superintelligence in Boston is hiring an ML infrastructure engineer to own and evolve the training and inference stack powering our scale-out research. You will design, implement, and operate distributed multi-GPU training jobs, model-serving pipelines, and the tooling researchers rely on to run experiments efficiently.

You will collaborate with engineering on capacity planning, observability, and reliability, and you will be the go-to person when pipelines stall or workloads fail.