Work

AI evaluation and solution review

I write and check mathematics where correctness matters: problems and reference solutions that test AI models, and step-by-step solutions that students can trust.

Experience

  1. MindriftAI evaluation

    Computational STEM problems

    Designing computation-heavy STEM problems, with Python reference solutions and scoring rubrics, to evaluate language models.

  2. MindriftAI evaluation

    Scientific-computing task review

    Reviewing and authoring Python tasks in scientific computing for correctness, reproducibility and complete specifications.

  3. PhotomathDec 2021 – Jan 2025

    Worked-solution review

    Writing and reviewing step-by-step solutions in algebra, geometry and calculus, checking every intermediate step as well as the answer.

Tools

Mathematics and computing

  • Python
  • NumPy
  • SciPy
  • SymPy
  • Matplotlib
  • LaTeX

Web

  • HTML
  • CSS
  • JavaScript
  • Git

Have an evaluation task in mind?

Tell me about the subject area and the kind of problems or reviews you need.

Get in touch