Python AI Coding Agent Evaluation Engineer

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We help professionals discover specialized projects, build a focused portfolio, and grow practical experience in a rapidly developing field. Creating an OpenTrain account is free, and this opportunity is available remotely worldwide.

• Remote contractor opportunity
• Part-time engagement with 20+ hours per week
• Open to candidates worldwide
• English-language work

About AI Coding Agent Evaluation

AI training is the human side of building artificial intelligence. For coding agents, experienced engineers review model behavior, code, tool use, and execution outcomes to help determine whether an agent can complete software engineering tasks accurately and efficiently.

This work combines hands-on programming with structured evaluation of LLM-powered workflows and development environments. Your technical judgment can help identify successful strategies, failure points, and practical opportunities to improve how AI agents operate.

• Review real AI-generated software engineering behavior
• Assess technical correctness and efficiency
• Contribute practical feedback to AI development
• Work remotely in a growing AI training field

The Role

OpenTrain is recruiting a Python AI Coding Agent Evaluation Engineer to support the development and evaluation of AI agents. You will examine complete agentic coding trajectories and determine whether an agent understands and executes software engineering tasks correctly.

The role involves reviewing prompts, tool calls, intermediate actions, code modifications, execution results, and final outputs. You will use software engineering judgment to distinguish technically correct solutions from flawed or inefficient approaches.

• Role level: Intermediate
• Work type: Contractor and part-time
• Time requirement: 20+ hours per week
• Work location: Worldwide and fully remote

What You'll Do

You will evaluate how AI coding agents reason through and complete development tasks. Reviews will cover the full trajectory from the initial prompt through tool usage, code changes, execution, debugging, and the final result.

The work requires reading, debugging, testing, and validating Python code produced or modified by AI agents. You will also assess LLM-powered workflows and development environments through structured technical review.

• Review prompts, tool calls, intermediate actions, and final outputs
• Assess whether agents correctly understand and execute software engineering tasks
• Identify technical errors, inefficient methods, and improvement opportunities
• Read, debug, test, and validate Python code
• Evaluate code modifications, execution results, and technical outputs
• Use Claude Code, Codex, Cursor, or comparable AI coding tools
• Analyze how agent actions affect the final result

Required Qualifications

This opportunity is intended for experienced software engineers with strong Python development skills and practical experience working with AI or LLM-based systems. You should be comfortable interpreting agent behavior, tool usage, intermediate execution steps, code changes, and resulting outputs.

• At least 5 years of professional software engineering experience
• Strong hands-on Python software engineering expertise
• Practical AI or LLM engineering experience
• Experience building agents, agent loops, LLM-powered applications, data pipelines, or similar systems
• Regular experience using AI coding agents in software development workflows
• Ability to evaluate technical correctness, efficiency, and outcomes

Helpful Background

Experience building agentic systems, LLM-powered applications, data pipelines, or related AI engineering systems is particularly relevant. Familiarity with multiple coding agents is useful because the evaluation work may cover behavior across different development environments.

Strong candidates can explain why an agent's approach succeeds or fails and how its intermediate actions influence the final output.

• Experience with Claude Code, Codex, Cursor, or equivalent tools
• Background building agentic systems or agent loops
• Experience developing LLM-powered applications
• Experience working with data pipelines
• Ability to explain flawed, correct, or inefficient technical approaches

Why Work With OpenTrain

AI training and data-labeling work gives technical professionals a direct role in shaping how modern AI systems behave. OpenTrain connects you with specialized evaluation projects, supports practical portfolio development, and helps you build a focused freelance career in AI development.

You will work remotely with professionals across a global network while contributing software engineering judgment to cutting-edge AI systems.

• Build practical experience evaluating advanced AI systems
• Contribute to the development of coding agents and LLM workflows
• Work remotely with a global professional network
• Build a durable portfolio of AI development contributions