Kevin Li | McGill University

Kevin Li

Kevin Li

I am a first year PhD Candidate at McGill University’s ECE departement working on software security and agentic AI risks under the supervision of Prof Lili Wei.

My research focuses on understanding, evaluating, and mitigating risks in modern AI and software systems, especially agentic AI systems and security-sensitive software engineering workflows.

Research Interests

  • Agentic AI risk protection
  • AI security and governance
  • Code summarization under obfuscation
  • Android secret leakage detection

Selected Projects

Agentic AI Risk Protection Framework

A framework for identifying, scoring, and mapping risks in agentic AI systems. The goal is to help organizations understand which risks are most relevant to their system and trace risks back to root causes and workflow components.

Android Checked-in Secrets Detection

A study on detecting hard-coded cloud API keys and secrets in Android applications using static analysis and validation techniques.

Code Summarization under Obfuscation

A project studying how to generate meaningful code summaries when source code identifiers are obfuscated or information-poor.