I’m a security researcher working on multi-agent safety and security. I’m currently a SPAR Research Fellow, focusing on automated evaluations and behavioural discovery for multi-agent systems, mentored by William Anderson and Joss Oliver from the Cooperative AI Foundation. I recently studied how a multi-agent system’s architecture changes its exposure to indirect prompt injection.

Before this, I spent a decade working on the security of decentralised systems, covering consensus protocols, applied cryptography and game-theoretic security. I was a research scientist at Protocol Labs (on the security of the Filecoin consensus protocol) and Scroll, and I interned at Meta. I also founded a startup in decentralised AI.

I hold a PhD from UCL’s Information Security Group, where my thesis on decentralisation and trust in cryptocurrencies was supervised by Sarah Meiklejohn and George Danezis.

Decentralized systems Consensus protocols Applied cryptography Multi-agent systems AI safety and security Game-theoretic security

SPAR Fellow — Cooperative AI Foundation

September 2026 – present · Automated evaluations and behavioral discovery for multi-agent systems · mentored by William Anderson and Joss Oliver

Parental career break

June 2025 – present

Research Scientist — Scroll

July 2024 – May 2025 · Interoperability research

Founder — Nelm AI

Mar 2023 – July 2024 · Decentralized AI

Research Scientist — Protocol Labs

Jan 2020 – Feb 2023 · Filecoin consensus security

Research Intern — Meta

June – Sept 2019 · Consensus protocols, Diem

Research Assistant — University College London

Sept 2018 – Dec 2019 · Applied cryptography & distributed systems

MRes + PhD in Computer Science — University College London

2014 – 2019 · Supervised by Sarah Meiklejohn and George Danezis