Member of Technical Staff, Coding Research
- Pay: $8 to $9 / hour
- Contract: Full time
- Who can apply: Open worldwide
- Posted Sep 14, 2026
In short
This full-time remote job asks you to help measure and improve AI coding agents, the programs that write code. You will design tests, scoring rules and data that show how well these models work. It pays 8 to 9 USD per hour and may suit experienced engineers who enjoy research and careful analysis.
What you will do
- Design and manage ways to score coding agents, including task specifications, scoring methods and quality rules.
- Lead research projects that measure and improve how well coding models handle many kinds of software tasks.
- Build high-quality datasets and reference examples so that tests give reliable results.
- Study where models make mistakes and turn what you find into fixes for training and testing.
- Build tools and pipelines that support large-scale experiments, reviews and evaluations.
What you need to apply
- Strong software engineering skills, for example in Python or C++.
- At least 3 years of experience in software engineering, machine learning, AI research, evaluation or a related technical field.
- Experience designing, reviewing or checking technical tests, benchmarks or evaluation methods.
- Clear written and spoken communication, and comfort working when priorities change quickly.
Who can apply
The job is open worldwide, with no country restriction stated, so you may apply from any country if you meet the experience requirement of 3 years.
My tip
We suggest you share a real example of a test, benchmark or evaluation you built, and explain the results in simple words, since this job rewards clear thinking.