Experience
From moderation team lead to head of Trust & Safety
Head of Global Compliance and Trust & Safety
MedalLead compliance and Trust & Safety for a gaming UGC platform moderating billions of videos, images and text submissions a month: AI-assisted moderation, privacy (GDPR, CCPA, DSA), youth safety, and CSAM escalations with NCMEC and law enforcement. Manage the vendor budget across multiple AI moderation and transcription vendors.
Safety Manager
MedalLed the zero-to-one build of a platform-wide user reputation system, projected to save $1M+ a year while improving moderation efficiency and enforcement accuracy. Led the migration to AI-assisted moderation and scanning for visual and text content.
Program Manager, Trust & Safety
MedalBuilt out the Trust & Safety program for a gaming platform handling millions of items a day: enforcement guidance, escalation workflows for child safety and severe harm, and the company's partnership with NCMEC. Introduced AI-assisted detection with human oversight for nuanced and high-risk cases.
Program Manager, Community Operations
MedalBuilt Medal's first Trust & Safety team, alongside its community, social and support functions. Wrote the Community Standards and enforcement playbooks from scratch.
Trust & Safety Team Lead
CognizantManaged three team managers, each leading about 30 moderation and community operations staff, and sent urgent policy guidance across a site of more than 300 moderators. Promoted to Supervisor in 2018.
What I work on
- Policy and enforcement
- Moderation operations
- AI-assisted moderation
- Child and youth safety
- NCMEC and law enforcement
- Privacy: GDPR, CCPA, DSA
- Vendor governance
- Escalations and crisis response
Education
T&S Workbench
A free, open set of Trust & Safety tools I built
Tools for finding risks before launch, rehearsing incidents, measuring what matters, choosing vendors and writing clearer policy. The knowledge behind them is published as open content. Private by default: work is saved only in your browser, and AI runs only when you ask.
- 58 abuse risks
- 104 safeguards
- 33 legal obligations
- 43 crisis scenarios
- 36 metrics
- 48 DSA duties
Pick a tool
- Your assessmentOne picture of your program: a score, grades, what to do first.
- Abuse pre-mortemHow a product could be misused before launch.
- Incident tabletopRehearse a crisis: 43 scenarios with the law behind each.
- Metrics framework36 metrics with formulas and SQL.
- Vendor scorecardChoose a moderation vendor on evidence.
- Coverage radarYour radar fills in as you answer.
- Program maturityRate 8 areas, get a phased roadmap.
- Classifier evalPrecision, recall and where a classifier fails.
- Compliance readinessWhich DSA and COPPA duties apply, gaps ranked.
- AI assistantsStress-test a policy, draft notices, review appeals.
- Red team studio · WIPTest your own AI feature for harm, one step at a time.
- Learn · WIPTwo red-teaming guides and two glossaries.
A guided assessment that fills in one picture of a Trust & Safety program: a score, each part graded, three radars and what to do first. Shown with the demo company.
Try itHow it worksSee how a product could be misused before launch, and what to do first.
Try itHow it worksRehearse a crisis: 43 scenarios, with a lesson and the law behind every call.
Try itHow it worksA reference to learn from: 36 metrics with formulas, SQL and how to measure each one.
Try itHow it worksChoose a moderation vendor on evidence, with rubrics and RFP questions.
Try itHow it worksOne question at a time, your radar filling in as you answer, then the gaps to close first.
Try itHow it worksRate a Trust & Safety program in 8 areas and get a phased roadmap for the biggest gaps.
Try itHow it worksDoes your moderation classifier apply the rule the way you meant it? Precision, recall and where it fails, by kind of case.
Try itHow it worksWhich DSA and COPPA duties apply to a service, article by article, with the gaps ranked and drafts for Legal.
Try itHow it worksStress-test a policy, draft enforcement notices and review appeals with AI, with a person making every call.
Try itHow it worksWork in progress. Test your own AI feature for harm, one step at a time: a target card, a try graded against an expert rubric, and a finding written up for an engineer to fix. A small team with no safety person gets a graded first finding in about twenty minutes, with a fix that matches the move tested.
Try itHow it worksWork in progress. Two guides to red teaming AI models, a red-teaming glossary and a plain-English Trust & Safety glossary, each with practice and progress saved in your browser. Method in cards, a check question per step, and 103 plain-word T&S terms checked against primary sources.
Try itHow it worksHandbook and writing
How I think about building and running Trust & Safety
I write about child safety, age assurance, automation and platform compliance, and I'm turning it into a free, open handbook on building and running a safety team.
The T&S Handbook · 19 chapters
From the first hire to a regulated program at scale
19 chapters in 4 parts, each paired with a free Workbench tool and built on ten principles from my writing.
Updated Oct 26, 2026
What changed
Recent writing
Mexico's age-check decree: build to the strictest bar you face
Mexico's presidential decree, in force since October 14, 2026, requires platforms, video games and app stores included, to use age checks that go beyond self-declaration, start new accounts on the strictest settings and give parents supervision tools, with a report due within 30 days. With no two markets asking for quite the same thing, build one approach to age assurance and child defaults to the strictest bar you face.
A T&S policy gets tested every day, on cases nobody wrote into it
A policy only works once it's built into reviewer guidance, product requirements and enforcement tools, and it's never finished: the edge cases reviewers hit every week should flow back into regular updates, each with a recorded reason. The test is consistency, the same call on the same case whoever reviews it and whoever's account it is, which the EU's Digital Services Act also expects.
How small can a T&S team get now that AI tools are this good?
AI tools now handle much of the clear-cut review volume, but regulators expect people who can own and explain decisions: qualified staff on appeals under the DSA and a named accountable person under Ofcom's codes. Automate the clear-cut volume and put the savings into senior owners of policy, appeals, risk and emerging threats.
Voice chat is where safety in games is furthest behind
In many games voice is only looked at after a player reports it, when the audio is already gone, so reports go nowhere and players stop reporting. The tools for proactive voice moderation exist and the law is catching up, so a game that puts children and adults in voice owes children the same protection there as in text.











