Shujaat Mirza

I want to understand how frontier AI can fail in ways we don’t expect, and to catch those failures before they happen. On Microsoft’s AI Red Team, I build attacks and evaluations to investigate emerging risks, with a focus on autonomy, loss of control, and multi-agent security.

Previously, I was a Research Manager at MATS. I completed my PhD at NYU’s Center for Cyber Security, advised by Dr. Christina Pöpper. I also worked on privacy-preserving machine learning at Spotify Research.

Interests

Autonomy and Loss of Control
Multi-agent Security
Monitoring and Oversight
Adversarial Machine Learning
Privacy-Enhancing Technologies

Education

PhD in Computer Science New York University

MPhil in Computer Science NYU Courant Institute

BS in Computer Science NYU Abu Dhabi

Recent Highlights

Oct 2026 Serving on the program committees for USENIX Security 2027 and AAAI 2027.

Aug 2026 Can a model's reasoning look benign while hiding a backdoor? In Evading Chain-of-Thought Monitoring Through Model Poisoning, we show how poisoned reasoning models can conceal backdoor behavior from their chain-of-thought traces.

Aug 2026 Presented "Red Teaming Agentic Networks" at Microsoft's AI & ML Connected Community.

Apr 2026 "None of us is as dangerous as all of us." We explored what breaks when AI agents become each other's infrastructure. Read the blog post →

Aug 2025 Distinguished Artifact Award at USENIX Security 2025 for reverse-engineering safety filters in DALL·E.