This five-week remote role involves turning real-world engineering workflows into terminal-based tasks, containerized environments, and automated grading tests to train and evaluate AI agents. The position seeks candidates with a Ph.D., postdoctoral background, or equivalent advanced engineering experience skilled in scientific programming (such as Python, C/C++, Julia, or MATLAB) and Linux. The work is conducted in English and requires at least 20 hours per week with four hours of overlap with Pacific Time.