Job Title: Software Engineer, AI Coding Tasks
Company: Askable, the evidence layer for the AI-powered enterprise
Job Type: Contract (minimum of one task a week, about 10 hours a week)
Location: Remote
Salary/Pay: US$1,000 per task that passes review, paid directly to you, with no limit on the number of tasks
Experience: 2+ years of professional software engineering. Code merged into a public open-source repository, ideally one you know well (a private codebase you own may also qualify, but not employer or client code). Education: Not specified
Skills Required:
- Programming: strong in at least one of Python, TypeScript, Go, Rust, Java, or C++
- Tooling: solid with Docker, Linux, Git, and automated testing
- AI tools: regular use of an AI coding tool
- Evaluation: experience writing tests, rubrics, or evaluation criteria, ideally for AI models
- Communication: ability to explain a change clearly and concisely in your own words
- Availability: at least 10 hours a week, starting now
- Equipment: a secure personal or work computer (no public or shared machines)
Advantage: Deep expertise in one area, such as databases, compilers, distributed systems, security, or infrastructure; experience building agent tasks or RL environments
Date Posted: Oct 20, 2026
Application Deadline: Not specified
Description:
Askable is hiring experienced software engineers to build AI coding tasks from a codebase they know well. Each task is based on a real change from a repository you know and must be hard for today's best AI coding models but fair for a skilled engineer. These tasks measure and train AI coding models. The process runs from application to proposal, build, stress-test, model testing, and review, with most tasks needing more than one round of revisions. Your work will include:
- Applying, completing a short screener interview, and uploading your resume
- Proposing a change from a repository you know and submitting a short proposal
- Setting up the repository as it was before the change, then writing the brief, the correct solution, and the grading checks
- Stress-testing the task with lazy, incorrect, and shortcut solutions to make sure your checks catch them
- Running the task against leading AI models until strong models often fail but a skilled engineer can clearly solve it
- Revising the task after review, then starting the next task once it is accepted
How to Apply: Apply here