Responsible AI Institute

Turning governance research into working frameworks

As a Research Engineer in AI Policy and Governance at the Responsible AI Institute (RAI), I build the theoretical and technical foundations for practical AI governance mechanisms. Recent projects include developing a risk classification framework for agentic AI, creating a policy-to-control mapping system, and establishing the foundations of an open source AI registry.

Role: Research Engineer in AI Policy and Governance
Since: January 2026
Work: ARC, TrustX, ROAR
Leading: The upcoming RAI Fellowship

How the frameworks connect

Hover a node · select a diamond to read more

ARC is grounded in established frameworks. TrustX builds off ARC, and ROAR maps AI applications to TrustX.

ARC · Risk classification
TrustX Agent Risk Classification (ARC)
ARC is a structured, iterable framework that provides risk classification for 7 agentic AI system types. It is grounded in established frameworks, such as the NIST AI RMF, the EU AI Act, ISO/IEC 42001, OWASP, MITRE ATLAS, and SR 11-7/26-2. Its outputs will then inform mapped control recommendations in RAI's policy generator tool.
Working paper (PDF)
7 system types12 risk dimensions
TrustX · Builds off ARC
TrustX Expanded Risk Framework
The expanded framework builds off of ARC by including procurement risk and exposure risk to legacy systems. It maintains the 12-risk dimension core introduced in ARC and tailors the risk classification to each risk surface with additional components, such as a procurement dossier or an agentic threat testing layer.
Working paper (PDF)
Procurement riskLegacy exposure
ROAR · Maps to TrustX
RAI Open AI Registry (ROAR)
A component of our launch of the TrustX for Finance Working Group was the introduction of ROAR, which allows individuals to contribute to an open source platform that shows how their AI applications map to our TrustX risk framework. This is a part of RAI's goals to create a collective core of iterative feedback within the AI governance community.
Sneak peek on the working group's site
Open sourceTrustX for Finance

RAI Fellowship

I'm leading the upcoming RAI Fellowship programme, which will mentor young graduates looking to break into the AI governance space.

Upcoming

Talks & panels

London Tech Week 2026 · Panel

Ensuring Growing AI Use Isn't Increasing Security Risk

Discussed the fluid yet nuanced definition of AI risk, the risks of shadow AI, how to manage risk at scale, and future steps towards effective AI governance.

AI Summit London 2026 · Partner-led meetup

Why 95% of Enterprise AI Projects Fail: The missing governance step that unlocks ROI

Facilitated a discussion about the real pain points organisations face with AI governance and how the Responsible AI Institute's TrustX framework and member community provide solutions and avenues for collaboration.

Slide deck (PDF)

How the frameworks connect

TrustX Agent Risk Classification (ARC)

TrustX Expanded Risk Framework

RAI Open AI Registry (ROAR)

RAI Fellowship

Talks & panels

Ensuring Growing AI Use Isn't Increasing Security Risk

Why 95% of Enterprise AI Projects Fail: The missing governance step that unlocks ROI

Policy & Insights newsletter