# Software Engineer, RL Data **Company:** [Cursor](https://scaleengineer.com/companies/cursor) Join Cursor's RL Data team to shape the training data for AI coding agents. You'll design and build systems that create high-quality, verifiable training data for reinforcement learning models, turning capability gaps into production-grade tasks. This role requires strong Python engineering fundamentals and expertise in data pipelines, evaluation infrastructure, and sandbox systems to ensure the data that powers the next generation of AI coding assistants is trustworthy and measurable. **Role:** Backend Engineer **Seniority:** Senior **Locations:** San Francisco **Salary:** 180000–280000 USD [Apply](https://jobs.ashbyhq.com/cursor/aef8f0e7-37cd-4d6e-96c9-3af72f588bae) Canonical: https://scaleengineer.com/jobs/cursor/software-engineer-rl-data-aef8f0e7 --- ## Responsibilities - Design and Scale Training Data Systems: Transform capability gaps into thousands of verifiable, production-grade tasks that train AI agents. Scale from hand-checked tasks to millions of data points using automated filters and evaluation systems while maintaining quality assurance. - Build Production-Grade Environments: Architect realistic environments that faithfully mirror production systems, including real codebases with dependencies, and simulations of applications like email, calendar, and docs that replicate actual behavior including search, error handling, and pagination patterns. - Develop Reward Mechanisms and Grading Systems: Design sophisticated graders for open-ended work that differentiate high-quality solutions from inferior ones without penalizing valid alternatives or inadvertently rewarding shortcuts and reward hacking behaviors. - Implement Quality Assurance and Validation Pipelines: Build automated systems to detect and prevent reward hacking, leaked answers, broken environments, and invalid execution paths before data reaches model training, ensuring data integrity at scale. - Conduct Data Analysis and Ablation Studies: Analyze training data effectiveness through rigorous experimentation, form hypotheses about data impact on model performance, run controlled ablations, and interpret rollout results to identify bugs, shortcuts, or data mismatches. - Collaborate on AI Research and Model Improvement: Work closely with research teams to understand how training data influences model capabilities and behavior. Translate findings into iterative improvements that measurably enhance the next generation of AI agents. ## Requirements ### education - {"name":"Computer Science or Related Field","description":"Bachelor's degree in Computer Science, Software Engineering, Mathematics, Physics, or equivalent professional experience demonstrating strong software engineering fundamentals."} ### technical - {"name":"Python Software Engineering","description":"Demonstrated expertise building non-trivial systems in Python including data pipelines, test infrastructure, evaluation frameworks, sandboxing solutions, or backend services at production scale."} - {"name":"Data Pipeline Architecture","description":"Experience designing and implementing scalable data pipelines that handle validation, transformation, and quality assurance of large datasets with automated checks and monitoring."} - {"name":"Test and Evaluation Infrastructure","description":"Strong background building test infrastructure, evaluation systems, or frameworks that ensure code quality and system reliability at scale."} - {"name":"Systems Engineering and Sandbox Design","description":"Experience architecting isolated execution environments, containerization, or sandboxing solutions that safely execute untrusted code while maintaining system integrity."} - {"name":"Metrics and Observability","description":"Proficiency in designing meaningful metrics, instrumentation, and observability systems to understand system behavior at aggregate and individual levels."} ### experience - {"name":"Systems-Level Software Development","description":"3+ years building backend systems, data infrastructure, or developer tools that process data at scale and require deep understanding of system reliability and performance."} - {"name":"Data Validation and Quality Systems","description":"Experience owning end-to-end data quality, validation pipelines, or automated testing systems that ensure correctness and prevent data corruption."} - {"name":"ML/AI Tooling or Infrastructure (Preferred)","description":"Background in machine learning operations, data engineering for ML systems, or building tools for AI development, though deep ML expertise is not required."} ## Skills ### required - {"name":"Python","description":"Production-level Python expertise for building scalable systems, data pipelines, and infrastructure code."} - {"name":"Systems Design and Architecture","description":"Ability to design robust systems that scale, handle edge cases, and maintain reliability under load."} - {"name":"Data Pipeline Development","description":"Experience building, debugging, and maintaining data processing pipelines with quality assurance mechanisms."} - {"name":"Problem Analysis and Debugging","description":"Systematic approach to identifying root causes through both aggregate metrics analysis and detailed examination of individual cases."} - {"name":"Software Testing and Validation","description":"Strong understanding of test design, quality assurance principles, and validation strategies for complex systems."} ### preferred - {"name":"Reinforcement Learning Fundamentals","description":"Understanding of RL concepts, reward functions, and training dynamics, though not required—strong engineers interested in learning are encouraged to apply."} - {"name":"Machine Learning Operations (MLOps)","description":"Experience with ML infrastructure, model training pipelines, or tools used in ML development workflows."} - {"name":"Containerization and Orchestration","description":"Familiarity with Docker, Kubernetes, or similar technologies for managing isolated execution environments."} - {"name":"Full-Stack Understanding of Code Execution","description":"Deep knowledge of how code executes, dependency management, environment simulation, and API behavior patterns."} - {"name":"Distributed Systems","description":"Experience building or operating distributed systems that scale to handle millions of tasks or events."} ## Tech stack ### tools - {"name":"Docker","description":"For containerizing isolated execution environments and ensuring reproducibility."} - {"name":"Kubernetes","description":"Potential orchestration platform for scaling task execution and data processing at scale."} - {"name":"Git and Version Control","description":"Standard version control for managing infrastructure code and experimental configurations."} - {"name":"Monitoring and Observability Tools","description":"Prometheus, Grafana, ELK stack, or similar for instrumenting systems and understanding behavior."} ### others - {"name":"Sandboxing Technologies","description":"Experience with isolation techniques, process sandboxes, or containerized execution for safe code evaluation."} - {"name":"API Testing and Simulation","description":"Building realistic simulations of external services (email, calendar, documentation APIs) for training environments."} ### databases - {"name":"PostgreSQL","description":"Recommended relational database for storing training task metadata and evaluation results."} - {"name":"Redis","description":"For caching, task queuing, and real-time metrics aggregation in data pipelines."} ### languages - {"name":"Python","description":"Primary language for building data pipelines, infrastructure, and evaluation systems."} - {"name":"SQL","description":"For data analysis, querying, and validation across training datasets."} ### frameworks - {"name":"FastAPI or Flask","description":"Common frameworks for building evaluation and task generation APIs."} - {"name":"PyTorch or TensorFlow","description":"Potential exposure to ML frameworks when interfacing with model training systems."} ## Benefits ### benefits - {"name":"Equity Stake","description":"Significant equity compensation as part of a high-growth AI infrastructure company backed by prominent venture capital firms."} - {"name":"Competitive Salary","description":"Market-competitive compensation for senior-level engineers in the San Francisco AI/ML sector."} - {"name":"Flat Organizational Structure","description":"Direct collaboration with company leadership, minimal bureaucracy, and significant influence over technical decisions and company direction."} - {"name":"Talent-Dense Team","description":"Work alongside exceptionally skilled engineers and researchers passionate about advancing AI capabilities and coding automation."} - {"name":"Cutting-Edge Technology","description":"Opportunity to work on frontier AI research problems and shape the infrastructure that powers the next generation of coding assistants."} - {"name":"Flexible Work Environment","description":"Collaborate with a company culture that values shipping code, spirited debate, and creative problem-solving."} ## Compensation - **max:** 280000 - **min:** 180000 - **currency:** USD - **stockOptions:** true ## Interview process ### steps - {"name":"Initial Application Screening","description":"Cursor reviews your resume and application to assess technical background, relevant experience with data systems, and alignment with the role's requirements."} - {"name":"Technical Phone Screens (2-3 Short Interviews)","description":"A series of brief technical assessments (typically 45-60 minutes each) conducted with team members. Expect questions covering system design, Python proficiency, data pipeline architecture, and problem-solving approaches. These are conversational and collaborative."} - {"name":"Systems Design and Architecture Discussion","description":"Deep-dive conversation on how you would design training data systems, validate data quality at scale, and architect environments for AI agent evaluation."} - {"name":"Practical Problem-Solving Session","description":"Collaborative technical session where you work through a real problem related to data validation, pipeline design, or system architecture alongside the interviewer."} - {"name":"On-Site Interview at San Francisco Office","description":"Full-day visit including a small practical project aligned with the role (building part of a data system, designing a grading mechanism, or analyzing data quality), technical discussions, architecture design conversations, and meetings with core team members to assess cultural fit and collaboration style."} - {"name":"Project Presentation and Discussion","description":"Walk through your on-site project with the team, explain your architectural choices, discuss tradeoffs, and answer questions about your approach and decision-making process."} ## Full description Our mission is to automate coding. The first step in our journey is to build the best tool for professional programmers, using a combination of inventive research, design, and engineering. Our organization is very flat, and our team is small and talent dense. We particularly like people who are truth-seeking, passionate, and creative. We enjoy spirited debate, crazy ideas, and shipping code. # **Software Engineer, RL Data** ### About the Role The data behind a model shapes it more than almost anything else. On the RL Data team at SpaceXAI, you decide what that data is for our agents: which problems they practice, what world they practice in, and how they're judged. You'll take a weakness, like an agent that gives up on hard bugs or misuses a calendar API, and turn it into training data that fixes it, then show the next model is measurably better. You form hypotheses about why data is helping or misleading the model, build the systems to test them, run ablations, and read the rollouts to find the bug, shortcut or mismatch behind a result. Every flaw you catch, or miss, shows up in what the next Grok models learn. ### What you’ll work on * Turning a capability gap into thousands of verifiable tasks — for example, planting subtle bugs deep in a real project's dependencies for the agent to root-cause. * Building environments that behave like production: real codebases with their real dependencies, and simulations of apps like email, calendar, and docs that match the real ones down to search, errors, and pagination. * Designing graders for open-ended work that separate good solutions from bad without penalizing valid alternatives or rewarding shortcuts. * Building automated checks that catch reward hacking, leaked answers, broken environments, and paths the real product wouldn't allow, before data reaches training. * Scaling from a thousand hand-checked tasks to millions, with filters and judges you can trust in place of a person reading each one ### You may be a fit if * You have strong software engineering fundamentals, and you've built non-trivial systems in Python: data pipelines, test or eval infrastructure, sandboxes, or backend services. * You have taste for what makes an environment trustworthy: faithful enough that skills transfer, strict enough that shortcuts fail, and simple enough to debug when a result looks wrong. * You are careful and systematic — understanding both aggregate metrics and close reading of raw outputs. * You enjoy turning messy, open-ended problems into repeatable, high-quality data production systems. * RL or ML experience is a plus but not required. We welcome strong engineers who want to do research on what makes models better. ### Applying If there appears to be a fit, we'll reach out to schedule 2-3 short technicals. After, we'll schedule an onsite in our office, where you'll work on a small project, discuss ideas, and meet the team.