Full Stack Engineer, ML Research Tools

Cursor is building the next generation of AI-powered coding tools, starting with an IDE designed for professional developers. The company combines cutting-edge research with thoughtful product design and engineering, operating as a lean, talent-focused team that values truth-seeking, creativity, and shipping working code.

Cursor is at the forefront of AI-driven development tools, leveraging machine learning and reinforcement learning to fundamentally change how programmers work. The organization prioritizes deep technical contributions from individuals who can move fast, think clearly, and collaborate across research and product teams.

You'll join the Engineering team as a Full Stack Engineer focused on ML Research Tools, working full-time from the San Francisco office. In this role, you'll design and build the infrastructure that researchers and external contributors rely on to create, review, submit, and monitor the reinforcement learning environments and tasks powering Cursor's training pipelines. This is a product-engineering position embedded directly within the research organization, where your work will meaningfully compress the time from task conception to production training data.

What you'll do

• Build end-to-end workflows for vendors and research teams to collaborate on task creation, submission, and acceptance into training infrastructure.
• Develop review and inspection tools that surface rollouts, transcripts, grader outputs, and quality signals, enabling informed decision-making about task viability.
• Create systems for environment health monitoring, failure analysis, versioning, and data cataloging that make training datasets intuitive to understand, manage, and expand.
• Design and implement self-serve interfaces using shared component libraries that let researchers and domain experts build and refine tasks with built-in quality validation.

What you'll bring

• Demonstrated experience shipping full-stack products where you owned systems spanning user interface through backend storage and services, using technologies like TypeScript, React, Node, Python, or Go.
• Expertise building data-dense tools such as transcript viewers, diff systems, review queues, observability platforms, or operational dashboards, with clear perspective on presenting structured data effectively.
• Track record establishing or maintaining design systems and component libraries that scale across growing teams.
• Experience designing review, QA, moderation, or acceptance workflows where you've had to anticipate and prevent users from circumventing checks.
• Genuine commitment to data quality and willingness to work directly with raw data; familiarity with evaluations, graders, or reinforcement learning systems is beneficial but not required.

Nice to have

• Background in reinforcement learning or data-quality infrastructure.
• Comfort operating in ambiguous environments and partnering with research scientists and subject matter experts.