Technical Product Manager, Trust & Safety (AI)

Location: Remote - United States

Salary: $180,000–$220,000 USD per year, based on experience, skills, and qualifications.

Company

hirify.global is a global data engineering company providing data, evaluation frameworks, platforms, services, and human expertise for organizations building and adopting Generative AI.

What you will do

• Define and ship data products for frontier labs, translating research goals into specifications for safety training data, reinforcement-learning environments, and evaluations.
• Design sandboxed environments with misalignment traps, cyber traps, integrity traps, injected instructions, exposed secrets, poisoned tools, and unsafe action scenarios.
• Specify safe-behavior rubrics, verifiers, and reward signals for agentic tasks, including appropriate refusal, escalation, and confirmation behavior.
• Own safety metrics such as attack success rate, over-refusal, trap trigger rate, and task completion under adversarial pressure; identify blind spots and drive iteration.
• Prioritize MVPs and technical and business tradeoffs, maintain a roadmap based on the threat landscape, and communicate product specifications and safety findings.
• Partner with lab researchers, ML and platform engineers, red-teamers, policy experts, and delivery leads managing expert annotator pools.

Requirements

• 5+ years of product management experience, including technical products in ML, AI data, security, or trust & safety.
• Hands-on familiarity with LLM training and evaluation, including SFT, RLHF/RLAIF, reward models, verifiers, and benchmark design.
• Working knowledge of agentic AI, including tool use, computer-use agents, multi-step planning, and related failure modes.
• Strong understanding of AI safety concepts such as reward hacking, deceptive alignment, jailbreaks, and prompt injection.
• Technical depth sufficient to read code, review environment specifications with engineers, and make sound technical tradeoffs.
• Experience designing metrics for risk, system performance, and user impact and using them for decision-making.
• Track record of shipping 0-to-1 products in fast-moving, ambiguous environments.
• Ability to explain complex technical ideas clearly to researchers, customers, and executives.
• Creative and adversarial thinking about how capable models may circumvent safeguards and existing checklists.

Nice to have

• Comfort with Python and sandboxed environments such as Docker, virtual machines, and browser automation.

Culture & Benefits

• Remote work restricted to the United States.
• The role focuses on responsible advancement of AI and building trustworthy systems at scale.
• Work involves collaboration across research, engineering, policy, red-team, and large-scale expert delivery functions.
• hirify.global has a 36+ year history delivering data and AI-related solutions.