Job description

ABOUT HANDSHAKE Handshake's mission is to organize expert human knowledge to advance the AI economy. Handshake AI works directly with frontier labs on their most consequential data, evaluation, and post-training challenges, building the systems that turn expert human knowledge into the data and evaluations that make frontier models better. You will work alongside engineers, researchers, operators, and builders from organizations including Scale AI, Meta, Google, Amazon, xAI, Notion, and Palantir—and help build the systems that make expert human knowledge useful for advancing AI. THE ROLE We are hiring a Member of Technical Staff, Evals to help define how frontier AI systems are measured, understood, and improved. This is a broad, high-ownership role for researchers who build. You will partner with AI researchers, domain experts, and customers to develop new benchmarks, reward and verifier systems, agent-evaluation methodologies, and data-quality techniques. You will work on questions at the center of frontier AI progress: what should be measured, how to design evaluations that reflect real capability, how to create high-signal feedback, and how to build the environments and data systems that make those answers actionable. Early members of the team will have unusual influence over our technical direction, operating culture, and the open-source software, benchmarks, and research products we build. We care more about demonstrated research capability, technical judgment, and a builder's mindset than a specific title, degree, or career path. Location: San Francisco preferred; we are open to exceptional candidates in other locations. WHAT YOU'LL DO - Design and build evaluation frameworks, benchmarks, and methodologies for frontier LLMs, AI agents, multimodal models, and reinforcement-learning environments. - Develop reward models, programmatic verifiers, graders, and other feedback systems that make model behavior measurable and improvable. - Research what makes evaluations representative, difficult, reliable, and resistant to shortcutting or reward hacking. - Build systems for high-quality human data, including expert task design, annotation methodologies, data-quality signals, and data-attribution techniques. - Run fast, rigorous iteration loops: prototype, evaluate, interpret results, diagnose failure modes, and turn learnings into the next benchmark or system. - Publicly contribute to the field through benchmarks, open-source tools, research, and technical writing. WHAT WE'RE LOOKING FOR - PhD in ML/AI, computer science, data science, or related fields (or equivalent research experience in industry). - Publications at top AI/ML venues like NeurIPS, ICML, ICLR, COLM. - Builders who enjoy tinkering with agents and shipping high-quality software, benchmarks, or datasets (e.g. a strong GitHub profile / OSS contributions, or product portfolio). - Strong Python skills, experience building scalable software, working with agents. - Strong knowledge of frontier AI: benchmarks, eval techniques, agent harnesses, post-training recipes, data shapes. - Comfort operating in an ambiguous, fast-moving environment with substantial ownership. WHY JOIN - Work at the very frontier of AI with most major AI labs, researching some of the most important problems in Data and Evaluations. - Publish results and work in public through open-source benchmarks and software, papers, and blogs. - Join a rapidly growing company whose data business grew from zero to nearly $1B run rate in a year. - Help build an early technical organization where your work shapes the roadmap, standards, and culture. - Attend (and publish at) conferences like NeurIPS, ICML, ICLR, COLM. PERKS Handshake delivers benefits that help you feel supported—and thrive at work and in life. The below benefits are for full-time US employees. 🎯 Ownership: Equity in a fast-growing company 💰 Financial Wellness: 401(k) match, competitive compensation, financial coaching 🍼 Family Support: Paid parental leave, fertility benefits, parental coaching 💝 Wellbeing: Medical, dental, and vision, mental health support, $500 wellness stipend 📚 Growth: $2,000 learning stipend, ongoing development 💻 Office: Commuting support, free lunch, and gym in our SF office 🏝 Time Off: Flexible PTO, 15 holidays + 2 flex days 🤝 Connection: Team outings & referral bonuses

Requirements

Must-haves

We are hiring a Member of Technical Staff, Evals to help define how frontier AI systems are measured, understood, and improved. This is a broad, high-ownership role for researchers who build.
Design and build evaluation frameworks, benchmarks, and methodologies for frontier LLMs, AI agents, multimodal models, and reinforcement-learning environments.
Develop reward models, programmatic verifiers, graders, and other feedback systems that make model behavior measurable and improvable.
Research what makes evaluations representative, difficult, reliable, and resistant to shortcutting or reward hacking.
Build systems for high-quality human data, including expert task design, annotation methodologies, data-quality signals, and data-attribution techniques.
Run fast, rigorous iteration loops: prototype, evaluate, interpret results, diagnose failure modes, and turn learnings into the next benchmark or system.
Publicly contribute to the field through benchmarks, open-source tools, research, and technical writing.
PhD in ML/AI, computer science, data science, or related fields (or equivalent research experience in industry).
Publications at top AI/ML venues like NeurIPS, ICML, ICLR, COLM.
Builders who enjoy tinkering with agents and shipping high-quality software, benchmarks, or datasets (e.g. a strong GitHub profile / OSS contributions, or product portfolio).
Strong Python skills, experience building scalable software, working with agents.
Strong knowledge of frontier AI: benchmarks, eval techniques, agent harnesses, post-training recipes, data shapes.
Comfort operating in an ambiguous, fast-moving environment with substantial ownership.
Work at the very frontier of AI with most major AI labs, researching some of the most important problems in Data and Evaluations.
Publish results and work in public through open-source benchmarks and software, papers, and blogs.
Join a rapidly growing company whose data business grew from zero to nearly $1B run rate in a year.
Help build an early technical organization where your work shapes the roadmap, standards, and culture.
Attend (and publish at) conferences like NeurIPS, ICML, ICLR, COLM.

Nice-to-haves

Location: San Francisco preferred; we are open to exceptional candidates in other locations.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aimlpython

Platform: ashby
Effort: quick

Benefits

Benefits detected in the listing.

Dental

Vision

401(K)

Pto

Parental Leave

Wellness