Full-Stack Software Engineer, Reinforcement Learning

Role overview Build the product and internal systems that help research teams and external partners create, assess, and improve reinforcement learning data. The work spans user-facing tools, backend services, and operational dashboards, with close collaboration across research, operations, and commercial teams.

Responsibilities - Develop tools for browsing environments, inspecting trajectories, reviewing task quality, and understanding model behavior. - Build partner workflows for creating, submitting, testing, and refining environments and training data. - Create dashboards that show environment quality, evaluation results, data collection progress, and pipeline health. - Design services and APIs connecting task authoring, data collection, evaluation, quality review, and training infrastructure. - Turn cross-functional needs into polished product improvements and ship them quickly.

Requirements - At least 3 years of full-stack experience building and shipping production systems, with technical study or equivalent fundamentals. - Proficiency in Python and a modern web stack such as React, TypeScript, or Next.js. - Practical experience with AI, machine learning, or reinforcement learning, and ownership of products from design through deployment. - Experience with backend services, APIs, and databases, as well as dashboards or monitoring interfaces. - Comfort working with cloud infrastructure, Docker, CI/CD, and production debugging. - Strong product judgment, independent execution, and communication with technical and non-technical teams.

Nice to have - Experience with data inspection, quality review, collection, labeling, or evaluation tools. - Familiarity with AWS, Kubernetes, Terraform, or Grafana.

Benefits and work setup - Annual base salary of $130,000–$225,000 USD. - Visa sponsorship and relocation support are available for qualified full-time candidates moving to the United States or Singapore. - The role is based on-site in San Francisco. Singapore-based and remote candidates with substantial overlap with San Francisco or Singapore time zones may also be considered.