Website Crossing Hurdles
Crossing Hurdles is a company building tools to evaluate and improve AI coding agents – basically, we’re testing how well AI can write and change software code. We’re looking for a Software Developer to help us create realistic coding challenges for these AI systems.
You’ll be building these challenges based on actual changes made to open-source projects, like fixing bugs or updating code. This involves working with a system called Harbor to run these tasks inside secure environments. A big part of the role is writing very clear instructions for the AI, specifying exactly what needs to be done. You’ll also write Python scripts to automatically check if the AI’s code actually works correctly. You’ll need to think about how to break down big coding tasks into smaller pieces that the AI can handle. Then, you’ll test everything thoroughly and make improvements to the challenges.
We’re looking for someone comfortable with both Python and JavaScript. Experience with AI coding benchmarks like SWE-bench is a plus, but not essential. It’s important you’re able to understand and navigate large codebases – experience with projects like Django, Flask, Node.js or similar would be really helpful. You should also be familiar with Git for version control and know your way around Docker containers. Writing tests is important too, so experience with testing frameworks like pytest is good.
This is a short-term contract role, paying $15 per hour. You can work from anywhere with a reliable internet connection – it’s fully remote, but we do need some overlap with Pacific Standard Time (PST), around 4 hours per day. We expect the work to be between 20-40 hours per week for 4 weeks.
This role would be a great fit for someone who enjoys a challenge and is interested in the cutting edge of AI and software development.
To apply for this job please visit bebee.com.