Guo's long-term research goal is to design effective and trustworthy machine-learning solutions for a wide range of security problems. His recent research includes designing foundation models for software and network security problems, building reinforcement learning-driven planning and scheduling systems for security problems, and improving the explainability and robustness of large models and reinforcement learning.