Senior Backend Engineer (Python/FastAPI) for AI Evaluation

You will help create and assess coding examples used to train and benchmark large language models. The work focuses on Python across backend services, data pipelines, and machine learning infrastructure, with additional coding in JavaScript, ReactJS, C, C++, Java, Rust, and Go.

• Curate code examples, develop solutions, correct errors, and assess AI-generated code for efficiency, scalability, reliability, and maintainability.
• Build Python agents and automated tools that verify solutions to software engineering tasks.
• Look for recurring errors and design ways to check the quality of proposed solutions.
• Work across prototyping, architecture and API design, production implementation, launch, experimentation, monitoring, and ongoing maintenance.

What it pays and takes

• Pay: Not specified.
• Time: Flexible commitment, with a requirement of 20 or more hours per week.
• Location: You must be based in the United States.
• Engagement: Contractor role, initially for one month, with possible extensions based on performance and fit.
• Requirements: At least three years of software engineering experience; strong Python skills with production-grade frameworks and tools; experience building full-stack applications and deploying scalable software.
• Requirements: Deep knowledge of software architecture, design, development, debugging, and code-quality assessment, plus clear written and spoken communication for explaining evaluations.
• A strong computer science foundation and experience with production systems at high-scale technology organizations are valuable. Practical engineering ability is prioritized over institutional background.

How it works

OpenTrain is a platform where people find AI training work and start an application. Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI training work

AI training work involves preparing and reviewing examples that help AI models learn, including code examples and model-generated code. Software engineering experience helps contributors assess those examples and build reliable ways to check model solutions.