Staff Software Engineer (AI Evals)

G2i; United States; Remote; USD 138,000 - 276,000; Full time

Job summary

G2i is seeking a Staff Software Engineer (AI Evals) to evaluate AI coding interactions without writing production code. You will assess whether outputs are useful, correct at a high level, and reflect strong engineering judgment.

Role emphasizes taste and feedback over syntax, requiring seniority in Go, Python, Rust, or TypeScript/JavaScript, with hands-on use of Codex, Claude Code, and Cursor. Start ASAP for ~20 hours/week at $100–$200/hour.

Qualifications

• Experience at staff/principal level or equivalent.
• Hands-on with Go, Python, Rust or TS/JS in modern AI-assisted dev workflows.
• Ability to give direct, opinionated feedback on engineering interactions.

Responsibilities

• Evaluate AI-generated coding interactions end-to-end.
• Judge usefulness, correctness, and engineering judgment of outputs.
• Provide clear feedback on what works, what doesn’t, and what’s misleading.
• Help define what great looks like when using Cursor and similar tools.

Skills

Go

Python

Rust

TypeScript/JavaScript

Tools

OpenAI Codex

Claude Code

Cursor

Job description

Staff Software Engineer (AI Evals)Contract | $100–$200/hour | 20-40 hrs/week | Start ASAP

We’re looking for highly experienced software engineer to help evaluate the quality of interactions with modern coding agents such as OpenAI Codex and Claude Code.

This is not a traditional engineering role.

You won’t be writing production code.You’ll be evaluating something harder: whether the model thinks like a great engineer.

What This Role Actually Is

You will assess how AI coding agents behave in real-world scenarios — focusing on:

• Whether the response makes sense

• Whether the preamble and reasoning are useful

• Whether the output reflects strong engineering judgment

• Whether the interaction feels right to an experienced developer

This role is about engineering taste — not syntax correctness.

What You’ll Be Doing

• Evaluate AI-generated coding interactions end-to-end

Judge whether outputs are:

• Useful
• Correct (at a high level)
• Aligned with how a strong engineer would think

• Assess the quality of explanations and reasoning, not just code

• Distinguish between different levels of response quality (e.g. what makes something a 2 vs 4)

Provide clear, opinionated feedback on:

• What worked
• What didn’t
• What felt “off” or misleading

• Help define what great looks like when interacting with tools like Cursor

What We Mean by “Taste”

We’re specifically looking for engineers who can answer questions like:

• Does this feel like something a strong engineer would actually say?

• Is this explanation helpful, or just technically correct?

• Is the model guiding the user well, or just dumping output?

• Would this interaction build or erode trust?

You should be comfortable making subjective but rigorous judgments.

Who You Are

• Staff / Principal-level engineer (or equivalent experience)

Strong background in one of the below:

• Go
• Python
• Rust
• TypeScript / JavaScript

Hands-on experience using:

• OpenAI Codex
• Claude Code
• Cursor

• Deep familiarity with modern AI-assisted dev workflows

• Able to evaluate code without needing to fully execute or deeply review every line

• Comfortable giving direct, opinionated feedback

• High bar for what “good engineering” looks like

Nice to Have

• Experience with tools like Cursor or similar AI-first IDEs

• Prior exposure to prompt design or evaluation workflows

• Experience mentoring senior engineers or defining engineering standards

Engagement Details

• Rate: $100–$200/hour

• Hours: ~20 hours+

• Duration: Through October w/possible extentions

• Start: ASAP

Process

• Take-home evaluation exercise
• Background Check
• Offer