Software Engineer, ML Research

The role

This role sits within Cursor's research effort, where engineers build the foundations for training and serving its coding models. The work happens side by side with researchers to shorten experiment cycles. The company describes a flat structure and a small, highly capable team.

What you would do

• Develop the distributed compute stack used for model training, serving, and reinforcement learning runs
• Create reusable software that makes large data processing jobs easier for researchers to launch
• Design pipelines that convert real usage data from Cursor's users into material for model training
• Partner closely with research staff so experiments can be repeated and iterated on quickly

What they asked for

• Solid background in distributed systems or infrastructure engineering
• Track record of owning projects from initial design through to production release
• Working understanding of how language models behave internally
• Genuine curiosity about machine learning topics