AI Model Policy Trainer, Mental Health

About Handshake

Handshake was founded on a simple belief that everyone deserves a path to a great career, regardless of where they went to school or who they know. Today, we power 25 million job seekers, 1 million+ employers, and 1,600 educational institutions.

In 2025, we started Handshake AI and built the fastest-growing AI data business in history. We work directly with frontier AI lab researchers to create evaluations, publish benchmarks, and push the boundary of data. We’ve grown from $0 to ~$1B run rate and pay ~$60M to over 30K individuals every month.

Why join Handshake now:

• Shape how every career evolves in the AI economy, at global scale, with impact your friends, family and peers can see and feel

• Partner hand-in-hand with world-class AI labs, Fortune 500 partners and the world’s top educational institutions

• Work together with engineers, scientists, operators, and more from Palantir, Meta, Scale AI, and former YC founders

• Build a massive, fast-growing business with billions in revenue

About Handshake AI

Human data is the core infrastructure to AI advancement. Frontier AI labs currently improve model capabilities with various data‑intensive post‑training techniques. We believe that data spend for AI training will increase by 3-5x in the next few years and continue for much longer as models take on new domains. Handshake AI supports all of the frontier AI labs, working on their most complex data at the largest scale.

As an AI Model Policy Trainer focused on mental health, you will help evaluate how conversational AI systems respond to people discussing emotional distress, mental‑health conditions, unusual beliefs, possible crises, and other sensitive experiences.

You will review conversations and model behavior against a detailed mental‑health safety policy. Your work will help determine whether an AI response is empathetic, appropriately grounded, and sensitive to risk without reinforcing unsupported or potentially harmful beliefs.

This work requires an understanding of why people turn to AI for conversation, reflection, emotional support, or help during difficult moments. Strong candidates will recognize that AI can be useful and accessible while also understanding risks such as inappropriate validation, missed signs of crisis, over reliance, misleading guidance, or reinforcement of harmful beliefs.

Direct experience supporting individuals or families through mental‑health or behavioral‑health challenges is required. This experience may come from clinical care, social work, nursing, crisis support, professional caregiving, certified peer support, family advocacy, care navigation, or another structured human‑services setting. AI or technology experience is helpful, but it does not substitute for this domain experience.

• Review user conversations and AI responses involving mental health, emotional distress, crisis, unusual beliefs, paranoia, hallucinations, mania, and related experiences

• Apply detailed policy taxonomies and evaluation criteria to classify user requests and model behavior

• Evaluate both individual messages and changes in risk or certainty across longer conversations

• Distinguish between validating a person’s emotions and validating an unsupported or potentially harmful explanation

• Assess whether AI responses remain neutral and reality‑based or improperly reinforce, elevate, or operationalize harmful beliefs

• Evaluate actions taken by AI systems through tools, including searches, messages, scheduling, purchases, file creation, and publication

• Write clear rationales that connect each evaluation decision to evidence from the conversation and policy

• Participate in calibration exercises and work with other trainers to improve consistency

• Identify ambiguous examples, gaps, and edge cases that may require policy clarification

• Review the work of other trainers and provide thoughtful, evidence‑based feedback

• Incorporate coaching and adapt as policies, rubrics, and model capabilities evolve

• Collaborate with researchers, policy specialists, operations teams, and other subject‑matter experts

• Work regularly with sensitive and potentially distressing content

Qu…