Software Engineer, AI Research and Evaluation (US-Based)

Create and assess coding datasets used to train and benchmark large language models. Use production engineering judgment to evaluate correctness, efficiency, scalability, reliability, and software quality.

• Curate code examples and develop precise solutions for training datasets.
• Correct and refine code in Python, C/C++, Rust, Go, Java, and JavaScript, including ReactJS.
• Assess AI-generated implementations for systems-level correctness, performance, scalability, and reliability.
• Build verification agents to identify error patterns and assess infrastructure code quality.
• Evaluate model capabilities across prototyping, architecture, API design, production implementation, monitoring, and operational maintenance.
• Design ways to automatically verify solutions to software engineering tasks and work with technical teams to improve coding results against performance benchmarks.

What it takes

• Strong expertise in systems programming, infrastructure, or backend development.
• Experience building and deploying scalable, production-grade software.
• Deep knowledge of software architecture, development, debugging, and code-quality review.
• Proficiency in one or more of Python, C/C++, Rust, Go, Java, or JavaScript.
• Ability to provide clear, structured written and spoken evaluation rationales.

Schedule and engagement

• Location: Open to candidates based in the United States.
• Schedule: Flexible, with a commitment of 20 or more hours per week.
• Engagement: Individual contractor, part-time.
• Initial term: One month, with potential extensions based on performance and fit.

How to apply

Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI training work

AI training work uses human-created examples and reviews to help improve artificial intelligence. In this role, your software engineering judgment helps shape how models write and assess code.