Founder & PhD Researcher · King's College London
Wendi KimberliSoto
Building governance systems that detect when AI agents drift from who they're supposed to be.
Adaptive AI education is advancing rapidly — modelling learner behaviour, personalising content in real time, building systems that respond to how people actually think. But this progress has been built almost exclusively for the normative population. The people who already have access, infrastructure, and institutional support.
Meanwhile, cybersecurity training for marginalised communities is expanding — more programs, more funding, more awareness. But the methods haven't kept up. The training is still static, linear, and built on assumptions about learning that were outdated a decade ago. It treats all learners the same and hopes repetition will be enough.
These two fields have never intersected. Adaptive systems don't account for structural disadvantage. Community-focused programs don't leverage what we now know about cognition. The result is a gap — and it's the people carrying the most risk who fall into it.
My PhD focuses on closing that gap.
About
Who I am and what I build
I am completing my PhD at King's College London, where my research focuses on adaptive cybersecurity education, recursive feedback architectures, and human-AI co-evolution. I build AI systems that model user cognition in real time — not to reduce people to data points, but to meet them where they are and make the change stick.
Independently, I built BridgetOS — a governance layer for agentic AI that detects behavioural drift and enforces identity stability. Two US provisional patents are pending on that work. My background spans clinical science, cybersecurity project management, and adaptive learning systems. The question underneath has never changed: how do complex systems actually change human behaviour?
PhD Research
Adaptive cybersecurity education, recursive feedback architectures, and human-AI co-evolution in training environments at King's College London
Cognitive AI
Real-time user cognition modeling that adapts to anxiety, cognitive load, and learning style to support durable behavioural change
Human-AI Co-Evolution
Designing systems where humans and AI develop together — not AI that replaces human thinking, but AI that strengthens it
AI Governance
Identity governance, drift detection, and behavioral stability in agentic AI — keeping humans thinking critically alongside AI systems
SonderSec
Trauma-informed cybersecurity training and tools for marginalised communities facing disproportionate digital risk
Currently
What I'm working on right now
In progress
PhD at King's College London
Finishing up my systematic literature review. Researching adaptive cybersecurity education, recursive feedback architectures, and human-AI co-evolution in training environments.
In progress
BridgetOS Patent Conversion
Converting the BridgetOS provisional patents to full utility / PCT filings — identity governance, drift detection, and behavioral stability in agentic AI systems.
In progress
O'Reilly Radar Essay
Finalizing an essay on agent identity drift and AI governance for O'Reilly Radar.
Journey
The path looks nonlinear. The question has been constant.
2 years, 450K followers
Life Confessions
An anonymous confessions platform on Tumblr where people submitted their secrets and I paired them with visuals. Posts routinely hit tens of thousands of reblogs and likes. 450K followers in two years — an early lesson in how people behave when given safe, anonymous space to be honest.
"Life is messy. And we're only human. We all do things we're ashamed of. We all do things we regret. We all carry burdens. We all have secrets. We all have dreams."
Early career
Clinical Science
Biochemistry, laboratory work, and vaccine research. Where the question of how systems change behaviour first took root.
MSc, 2024
The Human Firewall
MSc in Cybersecurity Project Management at Royal Holloway. Dissertation on strengthening cybersecurity in healthcare through project manager training. The bridge between clinical science and security — and the first time the research question became explicit.
Read thesis →Professional
Cybersecurity Project Management
Managing security initiatives at scale. Learned that the biggest vulnerabilities are human, not technical.
Founded
SonderSec
Nonprofit designing trauma-informed cybersecurity training for marginalised communities facing disproportionate digital risk.
Independent R&D
BridgetOS
Built a technical governance layer for agentic AI — identity governance, drift detection, and behavioral stability.
King's College London
PhD Research
Adaptive cybersecurity education, recursive feedback architectures, and human-AI co-evolution. Building an AI system that models user cognition in real time.
Writing
Essays on recursive systems, AI cognition, and the human loop
Behavioral Credentials: Why Static Authorization Fails Autonomous Agents
Apr 24, 2026 · 7 min read
Enterprise AI governance still authorizes agents as if they were stable software artifacts. They are not.
Read essayComparative Drift & Identity Analysis: Project Vend Phase 1 vs. Phase 2
Mar 2026 · 18 min read
Applies an identity-level drift analysis to Anthropic's Project Vend experiment, revealing that scaffolding fixes presentation but not disposition — the agent became a better-performing sycophant, not a fundamentally different decision-maker.
Coming soonAI Psychosis: The Danger of the Clean Mirror
Mar 1, 2026 · 4 min read
If the machine is a mirror, the greatest risk isn't that it shows you something ugly. It's that it shows you exactly what you want to see until you forget where you end and the latent space begins.
Coming soonEchoes of the Unstable Mind: How Recursive AI and Human Madness Spiral the Same Way
Jul 10, 2025 · 5 min read
We like to think machines are cold and humans are messy. That AI is logic, and madness is chaos. But what if they're built on the same engine?
Coming soonRecognition
Awards and conferences
IT Audit Scholarship
ISACA London Chapter · Jan 2026
King's Student AI Manifesto — Co-Signatory
Co-created institutional AI policy document shaping King's College London's approach to AI in education. Covers integrity, adaptability, ethics, and student-staff collaboration.
KCLSU · King's College London · Jan 2026 · Read manifesto
Upcoming
Infosecurity Europe 2026
Attended
Black Hat Europe 2025
BETT UK 2026
AI & Big Data Expo Global
Projects
What I've been building
BridgetOS is the commercial governance product. Aegis is the doctoral research prototype. SonderSec is the nonprofit arm. The Recursive Thinking Lab is a cognitive assessment platform built on the same drift detection engine.
Tech Stack
Technologies and tools I work with
Research
R&D
Framework Development
AI / ML
AI-Assisted Development
I design the systems and direct the architecture — AI handles the implementation.
Frontend
Backend
Databases
Auth & Security
DevOps
Stay in the loop
Occasional updates on AI governance research, BridgetOS development, and new writing. No spam.
Get In Touch
Let's connect
Interested in collaborating, discussing adaptive systems, or just want to say hi? I'd love to hear from you.
I'm open to research collaborations, speaking invitations, advisory roles in AI governance, and partnerships that serve underrepresented communities in cybersecurity.