Head of Safety

General Intuition & Medal; New York (NY); On-site; USD 180,000 - 260,000; Full time

Job summary

General Intuition & Medal is seeking a Head of Safety to own compliance end-to-end across content safety, user data, and platform transparency. You will design metrics, build dashboards, and lead the automation efforts that reduce manual review while maintaining high accuracy.

This role requires hands-on technical work, strong policy judgment, and the ability to navigate evolving regulations with the engineering team.

Qualifications

• Experience leading trust & safety or product compliance at consumer platforms.
• Automate manual review via classifiers, tooling, or vendor systems.
• Proficient in SQL; able to pull data and build eval harness.
• Ability to translate regulation into clear guidance and action.
• Strong communication and leadership to scale safety with automation.

Responsibilities

• Own safety function and shrink manual footprint through automation.
• Build and evaluate detection systems; set metrics and thresholds.
• Run cost-benefit analyses for policies, tools, and vendors.
• Lead roadmap to holistic compliance with engineering collaboration.
• Develop team, define internal principles and external voice.
• Stay ahead of GDPR/DSA and map obligations for the product.

Skills

SQL

Python

Trust & Safety

Content Policy

Product Compliance

Data-driven decisions

Automation

Metrics & dashboards

Job description

About The Company

General Intuition is the frontier lab for acting in space and time. We build large action models and world models that can perceive, predict, and act across virtual and physical environments. General Intuition builds on the strength of Medal, the world's largest and fastest-growing platform for gaming clips, where millions of gamers capture, share, and discover new games every year. We've raised over $670M from Khosla, GC, Valor, and Point72 since October 2025, and recently closed our latest round at a $6.2B valuation.

The Role

As Head of Safety, you own compliance end-to-end across content safety, user data, and platform transparency - from the moderation queue and the classifiers behind it to regulatory calls, product roadmap, and the systems that make the function smaller every quarter.

This is not a pure moderation role and not a policy-memo role. We are a lean company and intend to stay one: every hour of manual review, every vendor contract, and every classifier you tune is a cost, and you're expected to know what each one buys us. We need someone technical enough to build and test the systems that get us there, not just specify them - someone who instinctively reaches for the numbers before reaching for headcount.

What We're Looking For

• You own the safety function and shrink its manual footprint - improving the accuracy, automation, and cost-efficiency of content moderation until the same or better outcomes take materially less human review each quarter. You define the metrics behind it (queue volume, review time, appeal and reversal rates, cost per action, escalation latency) and build the dashboards that surface them without you in the loop.

• You build and evaluate detection systems hands-on - test sets, precision and recall, thresholds tuned against the real cost of false positives versus false negatives - and run the evals that tell you whether a change helped or just felt better. You've shipped changes to a detection or review system and can say how you measured them.

• You run cost-benefit on every action - before we invest in a policy, a tool, a vendor, or a review process, you can say what risk it reduces, what it costs to run, and what the cheaper alternative is. You reason in expected value: likelihood of harm, severity, cost to mitigate, cost to ignore. You can defend a decision to not do something, and you apply the same discipline to your own time, including knowing when to stop optimizing.

• You stay ahead of regulation and keep the compliance picture live - GDPR, the EU Digital Services Act, UK online safety and investigatory powers legislation, cyberbullying and minor-protection rules, and emerging regimes - translated into plain-language guidance the whole company can act on, and into an honest, current map of where the product meets its obligations, where the gaps are, what they cost to close, and what happens if we don't.

• You lead product work - driving the roadmap toward holistic compliance with a heavy bias to automation, so safety doesn't become a permanent cost center. You scope it with engineering as a peer.

• You build the team and set the position - developing existing teammates to own more of day-to-day operations, hiring where the math justifies it, defining Medal's internal principles and external voice on user safety in step with our brand and culture, and bringing research on real risk exposure so leadership sizes safety spend deliberately rather than reactively.

• You bring experience leading trust & safety, content policy, or product compliance at a consumer platform, ideally one with user-generated content at scale, plus a track record of automating manual review away through classifiers, tooling, vendor systems, or policy design that kills edge cases. You're comfortable in SQL and write enough Python to pull your own data, build an eval harness, and script a workflow - you don't file a ticket to find out how many items hit a queue last week. You know what classifiers and LLM-based review can and can't do, how to evaluate a vendor's claims, and where the failure modes are, without needing to train models yourself.

• You're direct - you look for the flaws in a plan, ask the skeptical question, and say the uncomfortable thing early rather than late, both as a manager and when reporting up, so hard decisions get made with good information.

This is a broad charter with room to grow: you'll be the company's center of gravity on digital trust and compliance, with direct access to leadership, engineering resources to build with, and the latitude to shape how Medal thinks about user safety as the regulatory landscape keeps shifting.