We're hiring Senior/Staff Full-Stack AI Engineers across three specialties. Each path is a real production problem we're solving inside Terramtech. Apply once — pick the challenge that plays to your strengths after you're accepted.
Build a system that extracts structured data from PDFs, CSVs, images, and free-text, then cross-validates and unifies it into a clean record.
Build an experiment tracking platform for a food R&D lab — clean noisy data, visualize patterns, and recommend the next experiments to run.
Build a computer vision pipeline that detects defects, measures dimensions against spec, and produces an operator quality report.
Transparent from day one. Senior engineers' time is expensive — we don't ask for a week of unpaid work.
Paid to every candidate who completes the full 7-day challenge, regardless of whether we extend an offer. We respect your time.
A hiring process designed to find builders, not talkers.
Submit your application with your GitHub username. No resume required — just your identity and a desire to build.
Once accepted, you get a private GitHub repo with a real-world engineering challenge. The clock starts ticking.
You have 7 days to architect, implement, test, and document your solution. Work how you actually work — we're watching the process, not just the output.
Our automated pipeline analyzes your code quality, security, architecture, and git discipline. These scores serve as one data point in a human-led evaluation process — no candidate is automatically rejected or advanced based solely on automated scoring.
Eight dimensions scored through automated analysis and expert review.
System design, separation of concerns, scalability, and design patterns
Prompt engineering, agent design, tool-calling, and error handling
Linting, type safety, complexity metrics, and clean code principles
Dependency scanning, secrets management, input validation, and auth
Infrastructure-as-code, deployment strategy, and observability
Coverage, test quality, edge cases, and testing strategy
Commit cadence, message quality, branching strategy, and PR workflow
Response to feedback, documentation quality, and collaboration signals
We analyze your entire commit history — cadence, message quality, refactoring patterns, and work habits. One giant commit of AI-generated code stands out immediately.
During the challenge, an AI-powered bot simulates a real engineering team — opening issues, reviewing PRs, and reporting bugs. How you respond reveals your collaboration skills.
Security scanners, linters, architecture analyzers, and test coverage tools provide objective measurements. Human reviewers add context and judgment.