Machine Learning Engineer — Cambridge, MA

Machine Learning Engineer | AI for Physics Research

Boston/Cambridge, MA | Hybrid - potential for fully remote

We’re partnering with a startup that has raised $60M in seed funding to build AI systems that investigate physical problems through reasoning, simulation and verification. Its first application tackles the power and cooling constraints behind AI data centers, with a longer-term ambition to accelerate fundamental scientific discovery.

The 27-person team brings experience from NVIDIA, Google, Carnegie Mellon, Harvard and MIT. You’ll join early enough to shape how researchers turn experiments into repeatable training runs and dependable services.

The role

You’ll connect research code to the wider engineering platform, taking ownership of the software and infrastructure that researchers use every day.

What you’ll own

• Turn research code into tested services with clear interfaces and predictable failure behavior.

• Build tools for launching experiments, routing model requests and integrating research workflows.

• Develop and operate infrastructure in your strongest area, including distributed training jobs, GPU scheduling, serving or agent systems.

• Investigate failures, resolve their causes and improve reliability.

• Work directly with researchers to make infrastructure easier to adopt while keeping operational tradeoffs clear.

What you’ll bring

• 3+ years operating production ML training or inference infrastructure.

• 3+ years engineering systems used by other people; these periods can overlap.

• Deep expertise in at least one of the four infrastructure areas.

• A concrete example of a failure you personally investigated and resolved.

• Strong software engineering judgment and comfort working closely with research teams.

Depth in one of the four ML infrastructure lanes required: training, serving, ML platform or agent/gateway/evaluation infrastructure.

Experience across a second infrastructure area, CUDA, ML performance optimization or cloud infrastructure would be valuable.

Please note: You will need to be either a US Citizen or Green Card Holder - no sponsorship available

Apply to find out more

Ethan Lewis - ••••••••