Senior Software Engineer, AI Training - US

G2i Inc.
•$100 - $200•Remote

About The Position

This is a contract role evaluating the quality of interactions with modern AI coding agents, including OpenAI Codex, Claude Code, and Cursor. It is not a traditional engineering role and does not involve writing production code. Instead, the work focuses on assessing whether AI coding agents respond and reason like strong engineers, including whether their explanations are useful, their outputs show sound engineering judgment, and their interactions would feel appropriate to experienced developers. The engagement is remote and open worldwide, although eligibility may vary by location. The role offers flexible hours of 10–20 per week and an hourly rate of $100–$200 depending on experience and location. The engagement is ongoing; individual projects may last from about two weeks to a few months, with new projects offered to successful evaluators. The start date depends on clearing a take-home exercise and an available project seat. The selection process consists of one take-home evaluation exercise with a recorded Loom walkthrough and no interview.

Requirements

  • Senior, Staff, or Principal-level engineer, or equivalent experience.
  • Strong background in TypeScript/JavaScript or Python.
  • Hands-on experience with at least one of OpenAI Codex, Claude Code, or Cursor.
  • Deep familiarity with modern AI-assisted development workflows.
  • Ability to evaluate code without executing it or reviewing every line.
  • Strong written and spoken English at B2 level or above; detailed written feedback and short recorded video explanations are required.
  • Comfort giving direct, opinionated feedback.
  • High standards for engineering quality.
  • Comfort making subjective but rigorous judgments and explaining them clearly.

Nice To Haves

  • Prior exposure to prompt design or evaluation workflows.
  • Experience mentoring senior engineers or defining engineering standards.

Responsibilities

  • Evaluate AI-generated coding interactions end to end.
  • Judge whether outputs are useful, correct at a high level, and aligned with how a strong engineer would think.
  • Assess the quality of explanations and reasoning, not just the code.
  • Distinguish between different levels of response quality.
  • Write clear, opinionated feedback explaining what worked, what did not, and what felt off or misleading.
  • Help define what great interactions look like when working with tools such as Cursor, Codex, and Claude Code.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service