Computational Chemist for AI Benchmark Design

Design graduate-level computational chemistry challenges that test whether advanced AI systems can use scientific software for research-style work, including running simulations, interpreting results, designing experiments, and extracting hidden information from data.

This role focuses on creating original benchmark problems grounded in real computational chemistry workflows. You will test problems with advanced AI models and refine them until they reach the intended level of difficulty. The work emphasizes thoughtful scientific reasoning, not routine data labeling or raw computational complexity.

• Create computational problems requiring skilled use of specialized scientific software.
• Develop fully specified, multi-step workflows that require an AI system to calculate exact answers correctly.
• Design open-ended problems in which the system must plan queries or experiments, interpret partial results, select useful measurements, and efficiently narrow possible explanations.
• Build problems involving simulations, result interpretation, experiment design, and information discovery from data.
• Test each problem against advanced AI models and revise the design to achieve the target difficulty.
• Write Python-based problem setups, oracle functions, and solution validators.

• Graduate-level training in a relevant STEM field, including an MS, PhD, or equivalent research experience. An MS or PhD is preferred.
• Deep hands-on experience with computational chemistry and electronic structure methods, especially PySCF.
• Experience performing quantum chemistry calculations such as Hartree-Fock, DFT, TDDFT, CASSCF, and post-HF methods.
• Ability to create problems involving excited-state analysis, orbital diagnostics, method selection for challenging electronic structures, and interpretation of artifacts caused by method limitations.
• Proven proficiency with at least one relevant scientific software library through research publications, open-source contributions, or professional work. Experience with other specialized software in this domain will also be considered.
• Strong Python skills and comfort working in Linux or terminal-based remote compute sandboxes.
• Ability to work independently, incorporate feedback, and refine problem designs.
• Strong scientific judgment, including an understanding of tool limitations, edge cases, and how to make a problem genuinely challenging through careful reasoning.

• Experience across multiple relevant scientific domains or tools.
• Familiarity with benchmark or evaluation design.
• Experience teaching scientific topics or designing exams and problem sets.
• Experience with computational reproducibility and containerized environments.

• Remote, hourly engagement.
• Availability of at least 15-20 hours per week is required.

$70-$100 per hour.