Astrophysics Problem Designer for AI Benchmarks

Obsidian; Houston (TX); On-site; USD 70,000 - 120,000; Part time

Job summary

Obsidian is seeking a Computational Astrophysics & Cosmology Expert to design graduate-level computational problems using real scientific workflows. You will craft tasks that require skilled use of software like Astropy, run simulations, interpret results, and guide experiments to reveal insights.

You will design problems that balance complexity and solvable pathways, working with AI models to test and refine difficulty while emphasizing reproducibility and thoughtful error analysis.

Qualifications

• Graduate-level training in a relevant STEM field (MS, PhD, or equivalent research experience).
• Proven proficiency with at least one of the listed scientific software libraries, shown through research publications, open-source contributions, or professional work.
• Strong Python skills — you'll be writing problem setups, oracle functions, and solution validators.
• Ability to work independently and refine problem designs based on feedback.
• Comfortable working in a Linux/terminal environment with remote compute sandboxes.
• Available for at least 15–20 hours per week.

Responsibilities

• Design original graduate-level problems based on real scientific workflows using specialized software.
• Test problems against state-of-the-art AI models and fine-tune difficulty until targets are met.
• Create problems where smart reasoning and strategic planning uncover hidden information, not just raw computation.

Skills

Python

Problem design thinking

Independent work

Education

MS/PhD in relevant STEM

Tools

Astropy

Cosmology libraries

Job description

Obsidian is seeking a Computational Astrophysics & Cosmology Expert to design graduate-level computational problems using real scientific workflows. You will craft tasks that require skilled use of software like Astropy, run simulations, interpret results, and guide experiments to reveal insights.

You will design problems that balance complexity and solvable pathways, working with AI models to test and refine difficulty while emphasizing reproducibility and thoughtful error analysis.