Materials Engineer & Python Expert – Freelance AI Trainer

About the role

Mindrift is looking for a Materials Engineer with strong Python expertise to design and author computational material‑science challenges for cutting‑edge AI models. This is a project‑based freelance position where you will create verifiable problems that require specialized scientific tools and evaluate AI performance in a sealed Linux container environment.

Key responsibilities

• Select a domain‑specific tool (e.g., ObsPy, MITgcm, flopy) and design a problem that hinges on its core algorithms.
• Develop a reference Python solution, provide all necessary input files and define the expected numerical answer with appropriate tolerance.
• Run parallel test batches, calibrate difficulty, and aim for a model pass rate of 10‑30%.
• Iterate on problem design based on feedback from senior reviewers in your sub‑field.
• Document the problem, solution, and evaluation criteria for future reuse.

Required profile

• Degree in Materials Science, Engineering, or a closely related discipline.
• At least 2 years of research, applied, or teaching experience.
• Proficient written English (C1 level or higher).
• Ability to work independently, learn new scientific packages quickly, and meet project deadlines.

Required skills

• Python programming.
• Experience with or willingness to learn at least one of the following packages: ObsPy, instaseis, pyrocko, MITgcm, xmitgcm, flopy, MODFLOW, GeoPandas.
• Familiarity with Linux container environments.

What we offer

• Project‑based freelance contracts with flexible remote work.
• Opportunity to influence AI model development in high‑impact scientific domains.
• Collaboration with leading tech companies and AI researchers.