Trustworthy AI-enabled software development

I focus on the reliability, safety, and trustworthiness of Code LLMs and AI-assisted development tools. Instead of evaluating AI-generated code only by whether it compiles or passes tests, my work asks whether the generated code is correct, safe, responsible, and aligned with developer intent. This direction includes harmfulness testing for Code LLMs, structured safety auditing of LLM-generated code, ethics testing for generative AI systems, and intent-centered validation of AI-generated software.