Head of AI Safety

Job description

Moonshot is recruiting a Head of AI Safety to lead the delivery, development, and growth of our AI Safety portfolio. The role combines Moonshot's expertise in violence prevention, behavioral risk, and online harms with the emerging practice of evaluating and improving the safety of AI systems. The portfolio addresses harm categories including pathways to violence, extremism, child sexual exploitation, abuse and grooming (CSEA), mental health and crisis, and risks affecting children and teenagers. The Head of AI Safety will serve as Moonshot's primary applied AI safety counterpart for frontier AI companies, governments, and regulators. The role will work closely with model, policy, trust and safety, product, research, and engineering teams. This is not an engineering or data-science role, but it is a hands-on position requiring the successful candidate to lead and participate directly in red teaming and adversarial evaluation, working in detail with evaluation methodologies, test scenarios, model responses, safety policies, and intervention frameworks. The role holds responsibility for client and partner relationships, project and staff management, methodological quality, and business development. The Head of AI Safety will build and maintain relationships across the wider AI safety ecosystem, including with governments, foundations, regulators, academics, researchers, and civil society organizations. Candidates must be based in MA, CO, NY, VA, GA, PA, MD, WI, TN, TX, OR, NJ, or DC. Occasional travel may be required.

Your responsibilities will include:

Applied AI Safety, Evaluation, and Advisory

Lead and quality-assure Moonshot's applied AI safety work across harm categories including pathways to violence, extremism, CSEA, abuse and grooming, mental health and crisis, and risks affecting children and teens, using methods such as red teaming and adversarial evaluation of AI systems

Advise frontier AI companies on how to improve the safety of their models, products, policies, and intervention systems

Translate insights from psychologists, child-safety specialists, violence-prevention practitioners, safeguarding experts, and other subject-matter experts into clear, actionable guidance for model safety, policy, product, research, and engineering teams

Set the methodological approach for the portfolio, translating violence-prevention, safeguarding, and behavioral-risk expertise into structured and testable evaluation frameworks

Lead and participate directly in red teaming and adversarial evaluation, working in detail with test scenarios, model responses, scoring criteria, safety policies, and evaluation results

Identify patterns, edge cases, and potential safety failures, and develop practical recommendations for improving model behavior and user protections

Maintain rigor and clear documentation across the team's technical deliverables, suitable for technical, government, and foundation audiences

Ensure work is delivered within a clear ethical framework and in compliance with contractual, legal, data protection, and ethics obligations

Identify, manage, and elevate operational, reputational, delivery, and partnership risks

Client & Partner Management

Serve as Moonshot's primary applied AI safety counterpart for frontier AI company partners, governments, regulators, and the wider ecosystem invested in AI safety

Build trusted relationships with model, policy, trust and safety, product, research, and engineering teams

Build and sustain relationships across the wider AI safety ecosystem, including governments, foundations, regulators, academics, researchers, civil society organizations, and specialist practitioners

Represent Moonshot externally in meetings, briefings, workshops, and sector engagement, including with regulators and policymaker audiences

Team Leadership & Management

Provide direct leadership, coaching, and management to Moonshot's AI safety team

Foster a collaborative, accountable, and mission-driven team culture, with particular attention to wellbeing given the sensitive nature of the work

Support workforce planning, performance management, and professional development across the team

Ensure effective coordination with internal teams supporting the portfolio, including operations, finance, research, and technical teams

Portfolio Development & Growth

Develop Moonshot's AI safety portfolio, identifying strategic opportunities, partnerships, and funding

Lead proposal development, scoping, and renewals with technical credibility, using precise, defensible language suited to technical and government audiences

Develop repeatable methodologies, service offerings, and partnerships that allow the portfolio to grow while maintaining methodological rigor and delivery quality

Support external communications, publications, briefings, and thought leadership that establish Moonshot as a credible voice in applied AI safety

Oversee project planning, staffing, budgeting, forecasting, and delivery timelines across the portfolio

Requirements

Essential:

Experience in trust & safety, online harms, or a closely related field such as violence prevention, safeguarding, or public health, and the ability to adapt that knowledge to AI systems

Curiosity about AI and the ability to build technical fluency quickly, enough to engage credibly with technical counterparts at AI companies. Much of this work is new, so comfort learning as you go matters more than existing AI safety expertise

Experience designing research, evaluation frameworks, or interventions for harm categories such as violent extremism, CSEA, self-harm and crisis, or targeted violence

Demonstrated experience managing projects, teams, budgets, partners, and clients, with strong people management skills

Excellent written communication, with experience producing credible (not promotional) material for government, foundation, or enterprise audiences

Comfort and demonstrated resilience working with highly sensitive or graphic content (CSEA, extremist material, crisis content), with awareness of wellbeing practices for this kind of work

Strong judgment and the ability to navigate ambiguity, competing priorities, and sensitive stakeholder environments, including representing organizations externally

Willingness to travel and work outside regular hours where needed to accommodate clients or respond to incidents

Highly trustworthy, with discretion and diplomacy, and willing to undertake relevant security clearance procedures

Experience supporting business development, grant funding, or procurement

Commitment to Moonshot's mission

Candidates must be eligible to work in the US, and will be required to undertake and pass any relevant security clearance procedures per client needs

Desirable:

Direct experience in model safety, red teaming, or adversarial evaluation of LLMs or other AI systems

Understanding of LLM architecture, safety tooling, or trust & safety policy

Prior experience in child safety evaluation, teen-safety product work, or grooming and CSEA detection

Familiarity with government or regulatory engagement, such as briefing officials or supporting policy submissions

Experience with intervention or diversion program design that can transfer to AI-mediated interventions

Academic or applied background in radicalization studies, forensic psychology, or violence risk assessment

Familiarity with taxonomy or classifier development, including how testing data feeds a classifier

Benefits

15 days paid vacation leave, plus Federal holidays and 1 day additional paid leave for Native American Heritage Day

Flexible public holiday policy with the option to work federal holidays in exchange for a day off at another time

Full private healthcare package, including coverage for partners and children

Dental & Vision Insurance

Life & Disability Insurance

24/7 access to free counseling via our Employee Assistance Program

3% matched 401k contributions

401(k) Roth Contributions

Generous maternity and paternity leave: 26 weeks paid maternity leave, 8 weeks paid paternity leave

All permanent employees are granted share options upon employment

Salary: $110,000 - $120,000 (depending on skills and experience).

#J-18808-Ljbffr

Worksite address

atlanta, GA, 30383, US

Who can apply

Review the original listing for work authorization, qualifications and employer requirements.