The work
- Design realistic, multi-step terminal tasks based on engineering and scientific workflows.
- Create simulation inputs, geometry files, sensor data, design constraints, configuration files, and related engineering datasets.
- Develop expert solutions using Python, C/C++, Julia, MATLAB or Octave, Bash, and relevant engineering software.
- Build containerized environments with appropriate tools and pinned dependencies.
- Create automated tests and objective grading criteria covering units, physical constraints, tolerances, convergence, stability, boundary conditions, and numerical behavior.
- Debug solver, dependency, workflow, precision, and performance issues.
- Document assumptions, expected outputs, and edge cases for each computational task.
What it pays and takes
- Pay: $300 per approved task.
- This is a remote contractor assignment expected to last 5 weeks.
- The commitment is at least 6 hours per day and a minimum of 40 hours per week, including 4 hours of overlap with Pacific Time.
- Open to candidates worldwide who are fluent in English.
- You should have a Ph.D., postdoctoral experience, or equivalent advanced technical experience in mechanical, electrical, chemical, aerospace, civil, materials, biomedical, robotics, control systems, or another engineering discipline.
- Strong scientific programming ability and hands-on Linux or terminal experience are required.
- Experience with engineering simulation, modeling, numerical analysis, optimization, signal processing, control systems, or technical data analysis is required.
- You should understand numerical methods, engineering units, physical constraints, boundary conditions, and technical validation.
- You should be able to build, debug, and validate reproducible computational engineering workflows.
- Helpful tools and background include NumPy, SciPy, pandas, matplotlib, SymPy, PyTorch, OpenFOAM, CalculiX, FEniCS, ROS, finite-element methods, computational fluid dynamics, robotics, embedded systems, digital twins, Docker, Git, CI/CD, automated testing, and HPC environments.
- Experience developing technical benchmarks, simulation-based evaluations, automated graders, or evaluations for AI coding and terminal agents is a plus. Publications, patents, open-source contributions, or industry experience involving computational engineering are also welcome.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training work uses human-created examples, reviews, and tests to help AI systems produce better results. Scientific computing experts are needed because they can judge whether code, simulations, numerical methods, and technical outputs are correct and reproducible.