Python Code Generation Reviewer

Your primary work is reviewing Python code produced by language models in response to natural-language coding prompts. For each task you'll receive the original prompt, one or more model-generated code snippets, and a rubric covering correctness, efficiency, readability, and edge-case handling. You'll run or mentally trace the code, identify bugs or logic errors, and submit structured feedback - including a preference ranking if multiple variants are presented and inline comments where issues occur.

Tasks arrive through a web-based review interface. You'll encounter a mix of prompt types: algorithmic challenges (sorting, graph traversal, dynamic programming), data manipulation with pandas or NumPy, API integration stubs, and general scripting tasks. Some sessions focus on ranking two completions; others ask you to rewrite a flawed snippet to a higher standard. You'll be working asynchronously and should expect 6–10 tasks per hour depending on code length and complexity.

Ideal candidates have 2+ years of Python development experience and are comfortable quickly reading unfamiliar code without an IDE scaffolding everything for you. You don't need deep ML knowledge - the job is software quality review, not model development. Familiarity with common Python anti-patterns, time complexity basics, and standard library modules (itertools, collections, os, etc.) will serve you well.

Skills

• Python proficiency
• Code correctness review
• Algorithmic complexity analysis
• Bug identification
• Inline code annotation
• Preference ranking

Where you can work

This role is 100% remote: you can work from anywhere in the world.