Mathematics Expert (India)

Pay: Up to $47/hour

Opportunity Overview

Handshake is looking for experienced mathematicians and research scientists to support AI research through flexible, part-time contract work. This is not a traditional job — you'll draw on your hands-on experience in pure or applied mathematics to evaluate AI-generated content and provide feedback that helps AI reason like a practicing mathematician.

This is an ongoing, project-based opportunity you can take on alongside your current role.

What You'll Do

• Use your domain expertise to create expert-level mathematics problems and scenarios grounded in real proof, formalization, and computational workflows

• Review and evaluate AI-generated responses for technical accuracy, rigor, and alignment with how the work is actually done

• Provide structured written feedback to help improve AI model performance

• Complete tasks at your own pace with no set schedule or minimum hour requirement

Qualifications

Ideal candidates meet all three of the following:

• Research experience — first-author publications and the ability to formalize proofs that AI models fail on

• Degree — PhD in Mathematics (pure or applied)

• Sub-domain depth — demonstrated expertise in one of the following areas:

  • Proof and formal methods (Lean, Coq, Isabelle)

  • Symbolic computation (SageMath, Mathematica), optimization, and numerical analysis

  • A core subfield such as algebra, topology, number theory, or combinatorics

Candidates who meet at least 2 of the 3 qualifications above are also encouraged to apply.

Additional qualifications (not required, but strengthen your application)

• Proficiency with theorem provers and symbolic computation software

• Experience formalizing proofs or working on problems at the edge of what current AI models can handle

Required for all applicants

• Must have Indian work authorization

• Strong written communication and attention to detail

• Ability to explain complex mathematical concepts clearly

• Ability to work independently in a remote, asynchronous environment