Secure Reasoning Lab

The Secure Reasoning Lab (SRL) is a research group in the Department of Computer Engineering at New York University Abu Dhabi, led by Prof. Jiahao Yu.

We work at the intersection of large language models, reinforcement learning, and cybersecurity. The lab builds AI systems that can reason about software security — finding vulnerabilities, generating patches, and remaining robust when those same models are attacked or misused.

Research Focus

• Cyber reasoning systems. Automated vulnerability discovery and program repair, including LLM agents for patch generation and ML-guided fuzzing.
• Secure and robust LLMs. Jailbreak evaluation, prompt-injection testing, and alignment methods that hold up under adversarial use.
• Reasoning and explainability. Improving, and explaining, the critical steps that LLM and RL systems take on security tasks.
• Blockchain security. Detecting anomalous transactions and securing decentralized systems with learning-based methods.

Collaborations

Prof. Yu co-founded the NYU Software Supply Chain Security Operations Center (NSOC) at NYU Tandon’s Center for Cybersecurity. NSOC embeds master’s students in open-source communities such as Python, the Linux kernel, and Rust for a year of real security work, including AI-assisted vulnerability discovery and patching, and publishes annual ecosystem assessments. See the NYU Tandon announcement.

News

We are recruiting self-motivated Ph.D. students and postdocs. If you are interested in large language models and cybersecurity, please see Join Us or email Prof. Yu.