STEM Researchers - Benchmark & First-Author Research

Gramian Consultancy is a boutique consultancy specializing in IT professional services and engineering talent solutions. With a strong background in software engineering and leadership, we help companies build high-performing teams by matching them with professionals who truly fit their needs.

Gramian Consultancy is a boutique consultancy specializing in IT professional services and engineering talent solutions. With a strong background in software engineering and leadership, we help companies build high-performing teams by matching them with professionals who truly fit their needs.

We're working with a highly specialized AI research lab building new benchmarks for how frontier AI systems perform real scientific research. They are looking for PhD students and postdocs across STEM fields to help reproduce papers, validate AI-generated research, and define what "good research" should look like for AI systems. We are looking for one Fellow per STEM field to help create a new benchmark for evaluating how well frontier AI models perform real scientific research.

The standout part of this opportunity: the benchmark will be released publicly with a research paper and open evaluation set, and Fellows will be named lead / first authors.

Research fellowship / contractor

Remote, global

Flexible hours

$30-$80/hour

Application → research review → interview

Biology, Medicine, Neuroscience, Materials Science, Chemistry, Physics, Mechanical Engineering, Chemical Engineering.

• Validate AI-generated paper reproductions, simulations, and research results.
• Map key research areas and taxonomy within your field.
• Define what "good research" looks like for AI systems.
• Create benchmark tasks and evaluation standards for frontier models.
• Contribute directly to the benchmark paper and public eval release

• Current PhD student or postdoc in a relevant STEM field.
• Based at a strong research university or institute.
• At least one published paper in your field.
• Able to critically review research and identify methodological or technical errors.
• Currently using AI for Science in your own research.
• Significant hands-on use of tools such as Claude Code, Codex, or similar AI agents

• Be a lead / first author on a public benchmark paper.
• Help create one of the first research benchmarks in your scientific field.
• Work directly with a niche AI research lab and frontier AI researchers.
• Flexible remote work designed around academic commitments