AI evaluation is not only a software-engineering problem. Strong evaluation work also depends on people who can recognize subtle mistakes in mathematics, science, finance, and domain-specific reasoning.
I organized nine technical specialties into a compact directory of Handshake AI opportunity guides. Each guide explains the kinds of reasoning the specialty can contribute and links to the corresponding official opportunity page.
1. Software engineering
Software engineers can evaluate implementation quality, debugging strategy, API design, testing, and whether a proposed fix actually satisfies a specification.
Software engineer opportunity guide
2. Machine learning
ML practitioners can inspect experimental design, model behavior, evaluation methodology, data leakage, and whether a conclusion is supported by the evidence.
Machine learning opportunity guide
3. Mathematics
Mathematicians are useful when a task depends on proof structure, edge cases, symbolic reasoning, or distinguishing a persuasive-looking argument from a valid one.
4. Physics
Physics expertise helps evaluate dimensional consistency, modeling assumptions, approximations, and whether a solution matches the behavior of the real system.
5. Chemistry
Chemists can catch errors involving reaction mechanisms, molecular structure, laboratory constraints, thermodynamics, and unsafe or unsupported conclusions.
6. Biology
Biologists can review causal claims, experimental interpretation, biological mechanisms, and whether an answer overgeneralizes from incomplete evidence.
7. Quantitative finance
Quantitative-finance researchers can evaluate statistical assumptions, time-series reasoning, risk models, backtesting logic, and hidden sources of look-ahead bias.
Quantitative finance opportunity guide
8. Investment banking
Investment-banking specialists can assess valuation logic, transaction mechanics, financial statements, market conventions, and whether a recommendation is grounded in the supplied facts.
Investment banking opportunity guide
9. Technical generalists
Some evaluation tasks reward breadth: breaking down unfamiliar problems, checking sources, identifying missing assumptions, and communicating a clear judgment.
Before applying
Treat every official opportunity page as the source of truth for current availability, requirements, selection, and compensation. Relevant expertise does not guarantee acceptance or project placement, and project availability can change.
Browse the complete Technical AI Opportunity Atlas
Referral disclosure
This is an independent guide, not an official Handshake website. The official opportunity buttons in the directory include my referral code. I may earn a referral bonus only if an eligible new fellow completes Handshake's qualifying paid production work. Applying does not guarantee acceptance, project placement, work, or payment.
Top comments (0)