AI Safety Engineer · Adversarial Testing · Detection Systems
Berenice Houston — I build real systems that keep kids safer online.
AI safety evaluation, adversarial testing and protective systems — focused on child safety and exploitation risks. I design, build and ship them end to end.
Selected work
Things I've shipped
Pick a case file.AI safety · Evaluation
ServalGuard®AI Evaluation
AI safety evaluation focused on child safety and exploitation risks — one product in the ServalGuard family.
AI safety evaluation and child-protection tools · 2023 to present
RavenTrak®
The studio behind it all — tracking, testing and hardening AI systems.
AI agents · Community support
Free AI bots for communities
Free AI bots for communities on Discord, Telegram, Reddit, and other platforms. Built for practical use and safety-related support. I developed these agents and made them available at no cost to communities.
Interactive safety game
ServalSim: Safi Detective
An interactive safety mystery starring Safi the Serval, built to help children and families recognize grooming risks.
Research · Tooling
Adversarial Lab
Prompt-attack suites for stress-testing models against misuse.
Nonprofit · Research
Center for Digital Harm Prevention
I created the Center for Digital Harm Prevention for nonprofit research. Beyond The Code explores why traditional AI guardrails fail to protect autistic users from online exploitation.
Developer · Apps
Diverse app concepts
I design and develop app concepts beyond AI safety, including a cooking app and a game for neurodivergent users.
2
registered trademarks
1
patent pending
1
person, every role
10+
projects built