Backend Developer Node.js (AI Evaluation)

Job Summary

As a Software Engineer, you will review AI-generated code and compare it with how you would write the solution yourself. Your main responsibility will be to evaluate AI-generated solutions, identify mistakes, and uncover reasoning gaps.

Right now, our developers are helping other engineers who work with AI get more accurate and reliable code. So, if you’ve ever written a prompt for AI and received a great answer, there’s a chance that work done by our developers helped make that answer better.

Requirements

3-4 years of hands-on experience as a Software Developer on commercial projects, including B2B SaaS systems and data-heavy applications.

Use AI tools (AI agents, Claude Code, Cursor, Codex or similar) in daily workflow.

Experience working with existing codebases and unfamiliar repositories

Strong experience with Node.js

Strong knowledge of JavaScript / TypeScript

Experience with backend frameworks such as Express.js or NestJS or Fastify

Good understanding of REST API and backend architecture

Experience working with databases (SQL and/or NoSQL)

Ability to independently analyze code and identify bugs or inconsistencies

Strong analytical and problem-solving skills

Upper-Intermediate English (B2)

Key Responsibilities

Work with Frontier Labs, producing training data to train Al models.

Evaluate AI-generated code, technical solutions, and engineering approaches.

Identify bugs, incorrect assumptions, architectural issues, and weak reasoning in AI responses.

Test and evaluate agentic coding models designed to perform complex engineering tasks at a senior developer level.

Compare AI-generated solutions against senior engineering standards and best practices.

Bonus:

Computer Science or related education.

Learning new programming languages.

Strong understanding of algorithms and data structures.

This is not a traditional software development role. You will use your engineering expertise to analyze AI outputs, find issues in their solutions, and help train better AI systems.