🚀 Hiring: Odyssey Task Author / Software Engineer - (Remote)
We are looking for an experienced Software Engineer / AI Coding Benchmark Task Author to create high-quality tasks for the Odyssey project on AfterQuery.
What you’ll need to do:
• Design a novel, realistic software-engineering task
• Choose the appropriate task/collection family
• Write a clear "instruction.md"
• Build the required Docker environment
• Create robust visible + hidden verification tests
• Develop a working reference/oracle solution
• Configure "task.toml"
• Test the task end-to-end
• Package the complete task in the required ZIP structure
• Ensure the task is difficult enough for a frontier coding agent but still objectively solvable
Required structure:
task/
├── task.toml
├── instruction.md
├── environment/
│ └── Dockerfile
├── tests/
│ └── test.sh
└── solution/
└── solve.sh
Ideal candidate:
• 2+ years of professional software engineering experience
• Strong experience with Python/JavaScript/TypeScript/C++ or similar
• Docker/Linux experience
• Strong understanding of automated testing and verification
• Experience with Terminal-Bench, Harbor, coding agents, SWE-bench, or similar projects is a big plus
• Able to create original, realistic, and non-gameable tasks
• Comfortable independently building and validating the complete task bundle
💰 Payment:
This is a pay-per-task opportunity. You will be paid for each task you complete and that is accepted after review. The more accepted tasks you complete, the more you earn. There is no fixed monthly salary—the compensation is based on the number of accepted tasks.
💬 Interested? Send your resume, relevant projects/GitHub, and your expected rate per completed task to [email protected]
Get in touch with us only if you can start immediately and can independently build, test, and validate the complete task bundle from start to finish.