Software Engineer - AI Coding Benchmarks & RL
Listing checked October 9, 2026 · pay as published by Terac
Overview
Terac builds coding tasks that train and test frontier AI coding agents. Experienced software engineers design realistic full-stack engineering problems, write tests that judge correctness, and push beyond the UI into services, APIs, and backend systems. The codebase for these tasks is TypeScript with React, Node, SQLite, and Playwright, and it grows as each task builds on the last. Contractors work 10 to 20+ hours per week, paid each week at $45 to $75 per hour, with US and Canada timezone overlap preferred. You can use AI tools as much as you like while writing code and tests.
What You'll Do8
- 1Craft full-stack web engineering problems that mirror real backlog items a product team would face.
- 2Build tasks on a codebase that expands with each new challenge, so later work depends on earlier decisions.
- 3Write test suites that determine whether a submitted solution works.
- 4Extend tasks past the browser UI into new services, APIs, and backend components.
- 5Implement backend logic in a second language such as Go, Rust, or Java.
- 6Add gRPC endpoints, rate limiting, and other system-level features drawn from production experience.
- 7Contribute observations from past projects and call out areas where the current task set should evolve.
- 8Use AI coding assistants during implementation and test writing.
Requirements8
- 1Professional software engineering background across both backend and frontend.
- 2Hands-on experience building and debugging production systems.
- 3Strong testing habits and a track record of writing reliable tests.
- 4Clear written communication and ability to work from written documentation with little supervision.
- 5Availability for 10 to 20+ hours per week, with a near-term start.
- 6Residence in the US or Canada preferred for timezone alignment; exceptional candidates from other regions considered.
- 7Active LinkedIn and GitHub profiles required.
- 8Willingness to complete a short screener, a brief AI interview, and a paid work trial on the codebase.
Who Should Apply
Software engineers with solid full-stack experience fit this role best. You have built and debugged real systems, care about testing, and can write tasks that mirror actual engineering work. The contract suits independent contributors who communicate in writing and can commit 10 to 20+ hours per week starting soon. Candidates who need heavy direction, dislike documentation, or lack backend depth beyond the UI tend to score lower, and rejection often follows from thin GitHub history, an inactive LinkedIn profile, or tests that only check surface behavior. Timezone overlap with US or Canada helps, though strong candidates elsewhere still get consideration.
Salary Insight
Terac pays $45 to $75 per hour for this contract, with funds sent each week. That range matches mid-to-senior full-stack contract work in the US and Canada, and the top end fits engineers who bring backend depth in Go, Rust, or Java plus strong testing habits.
Pay and demand for Software Engineering roles
AggregatedTypical pay
$70/hour
This role
Pay not disclosed
Most Software Engineering roles pay $44–$90 per hour.
Based on 304 similar roles that publish pay · 115 publish only a top rate; those count at the rate they gave
Rates shown per hour. Yearly and monthly pay converted; one-time fees and non-USD pay are not included.
- Live similar roles
- 363
- Listed in last 30 days
- 291
- Remote
- 94%
Hiring most right now: SME Careers (114) · micro1 (102) · Mercor (30)
Most requested skills · share of roles
- llm evaluation25%
- ai training24%
- python23%
- trainer feedback22%
Figures from Software Engineering roles live on NearSkill when this page loaded. A role can close before you apply, so check the listing itself.
Compare your resume against these rolesLocation
Compensation
Undisclosed by Terac
Comparable roles pay
304 roles$44 – $90/ hr
Not this role’s pay. Terac has not published a range.
Confirmed during Terac profile screening.
Required Skills
Application Tip
In your application, link a GitHub repo that shows full-stack tasks with real tests, and call out a backend component you wrote in Go, Rust, or Java. Pair that with a short note on how you debug and test systems, since Terac screens for testing rigor and written communication.
See NearSkill jobs more often in your search
Application & verification flow
1Instant rubric match
Your resume is scanned against this role’s requirements to check qualification fit.
2Screened before the employer sees it
Only profiles that clear screening are passed on.
3Outcome by email
We notify you at the address on your resume once the screening is reviewed.
Similar open positions
Explore active roles that match your skills and interests.
Micro1
VerifiedSoftware Engineer – AI Training
This remote contractor role at micro1 puts your backend coding skills to work training next-generation AI. You will design realistic coding challenges that involve feature implementation and bug fixes, then build deterministic verifiers that accept valid solutions and reject everything else. Expect about 15 hours a week, with a flexible schedule you control, including weekends if you want. No prior AI experience is needed. Your domain knowledge in languages like Java, Rust, Python, or TypeScript is what matters.
Micro1
VerifiedSenior Software Engineer
micro1 is hiring Senior Software Engineers to build Reinforcement Learning Environments that test how well AI models handle real coding problems. You will design tasks around bug fixing, feature implementation, refactoring, and performance tuning, then create a reproducible setup and a golden reference solution for each. This is a remote contractor role at roughly 15 hours per week, and no prior AI experience is needed. Your software engineering depth is what matters here.
Terac
VerifiedSoftware Engineer - AI Agent Code Review
Terac pays software engineers to review coding tasks and evaluation harnesses used for benchmarking AI agents. You will inspect programming challenges, determine whether they mirror real software engineering problems, and verify logic, test cases, and environment structure. The remote session combines a guided conversation with screen-sharing exercises, so you explain your reasoning as you analyze code. Engineers in the USA, Canada, and France can take part.
Micro1
VerifiedGitHub Contributor for AI Training (Remote, Contract)
micro1 is hiring expert Senior Software Engineers to help train next-generation AI systems by building Reinforcement Learning Environments. You will create reproducible coding challenges and golden reference solutions that test an AI model's ability to fix bugs, add features, refactor code, and improve performance. This is a remote contractor role, roughly 15 hours per week, with compensation of $50 to $100 per hour based on output. No prior AI experience is needed. Your deep software engineering skills in languages like Python, Java, Rust, Go, C++, or TypeScript are what matter.
Turing
VerifiedSoftware Engineer – AI Code Evaluation & Benchmarking (US candidates only)
This contract role places an experienced software engineer inside an evaluation workflow for frontier AI models. You will review AI-generated code for correctness, efficiency, and maintainability, validate solutions against real engineering tasks, and debug failures across different environments. You will also help build and refine evaluation datasets, benchmarks, and grading rubrics. The assignment runs for one month, requires at least 4 hours per day and 20 hours per week with a 4-hour overlap with PST, and is open only to candidates in the US. Work is fully remote.
Terac
VerifiedSoftware Engineer - AI Coding Evaluation (South Asia)
Terac wants experienced software engineers across South Asia to test the quality of environments used to score AI agents on coding work. You will join an AI-moderated session, inspect proposed programming tasks, and judge how well they match real engineering demands. Your notes will cover task structure, difficulty, and realism, plus where the coding evaluation harnesses need fixes. The project is remote, paid, and open to engineers in India and the Philippines.