Mathematics Researcher - AI Systems

Obsidian; New York (NY); On-site; USD 50,000 - 95,000; Part time

Job summary

Obsidian is recruiting mathematicians to contribute to AI systems focused on scientific reasoning. The role involves authoring mathematical problems and evaluating AI outputs alongside a collaborative team.

Successful candidates will hold a PhD and have a robust publication record. This part-time position emphasizes depth of research experience, appealing to those seeking to influence emerging AI technologies.

Qualifications

• First-author or sole-author publications in top mathematics journals.
• Doctoral or postdoctoral training from a top-20 global institution.
• Active research role as PhD candidate, postdoc, or faculty.

Responsibilities

• Author and review challenging mathematics problems for AI models.
• Evaluate AI model outputs for validity and accuracy.
• Collaborate with the program team on research tasks.

Skills

Research in mathematics

Problem-solving

Publication in top journals

Evaluation of AI models

Education

PhD in Mathematics or related field

Job description

We are recruiting a select cohort of accomplished mathematicians to contribute to the next generation of AI systems for scientific and mathematical reasoning. Our partner labs are building models that need to reason about pure and applied mathematics — number theory, algebra, algebraic geometry, topology, analysis, PDEs, probability, combinatorics, dynamical systems, logic, representation theory, mathematical physics, and adjacent areas of theoretical computer science and statistics — at the level of a working research mathematician, not a textbook. To get there, they need contributors whose own research sits at the frontier.

This is a W-2 employment position with Cincinnatus LLC, requiring a part-time commitment of up to 20 hours per week. This position will be placed at a leading AI lab as part of their extended workforce.

What you'll do

• Author and review challenging mathematics problems used to train and evaluate frontier AI models
• Evaluate model outputs on tasks ranging from proof construction and verification to problem-solving, formalization, and conjecture exploration
• Surface areas where current models fail — especially the subtle errors a working mathematician would catch immediately: an unjustified step, a misapplied hypothesis, a hand-waved lemma, or a proof that is plausible but wrong
• Collaborate with a small program team and other senior contributors

The level we're hiring at

This is not a generalist contributor pool. We're looking for researchers whose CV would plausibly include items like:

• First-author or sole-author publications in journals at the level of Annals of Mathematics, Inventiones Mathematicae, Journal of the AMS, Acta Mathematica, Duke Mathematical Journal, Communications on Pure and Applied Mathematics, Geometry & Topology, JEMS, GAFA, Advances in Mathematics, Probability Theory and Related Fields, or top specialty journals (for adjacent fields, top-tier venues such as STOC, FOCS, SODA, or Annals of Statistics also qualify)
• Doctoral or postdoctoral training at a top-20 global institution for mathematics or STEM
• Competitive fellowships and honors such as the Clay Research Fellowship, NSF GRFP / NSF Postdoctoral Fellowship, Sloan Research Fellowship, Packard Fellowship, Simons Fellowship or Simons Investigator, AMS Centennial Fellowship, Hertz Fellowship, Marshall / Rhodes Scholarship, New Horizons in Mathematics Prize, Salem Prize, Ramanujan Prize, or equivalent — and, for earlier-career applicants, signals such as Putnam Fellow placement or International Mathematical Olympiad medals
• NSF CAREER or other grants where the contributor was a named PI or co-PI
• Currently or recently in an active research role: PhD candidate, postdoc, instructor, faculty, principal researcher, or industry research mathematician

Outstanding senior PhD students and recent graduates with strong publication records are explicitly encouraged to apply — years of experience matter less than the depth of the research output.

Engagement

• Part-time: 8 to 20 hours per week
• Compensation reflects research seniority and credentials
• No fixed end date; contributors continue as long as the program needs their expertise

About Cincinnatus LLC

Cincinnatus LLC is an enterprise staffing company that partners with leading technology companies to source and employ highly skilled professionals for contingent and contract-based opportunities. Cincinnatus serves as the employer of record for these engagements, providing W-2 employment, payroll, benefits, and compliance, while placing employees directly within client teams to work on high-impact initiatives.

Equal Employment Opportunity

Cincinnatus is proud to be an Equal Employment Opportunity employer. We do not discriminate based upon race, religion, color, national origin, sex (including pregnancy, childbirth, reproductive health decisions, or related medical conditions), sexual orientation, gender identity, gender expression, age, status as a protected veteran, status as an individual with a disability, genetic information, political views or activity, or any other legally protected characteristic.