I’m a Research Assistant Professor in the Sanghani Center for Artificial Intelligence and Data Analytics at Virginia Tech. My background is in Computer Science, and I work at the intersection of software engineering and artificial intelligence. Prior to my current appointment, I was a postdoctoral associate at Virginia Tech. I earned my M.S. and Ph.D. in Computer Science from the University of Texas at Arlington under the advisement of Prof. Jeff Lei.

Research Interests

My research is at the intersection of software engineering and artificial intelligence. Over the past decade, my work has focused on addressing the multi-faceted challenges in testing AI-enabled systems across its lifecycle. The question that motivates my research is: How do we systematically test and evaluate AI systems to ensure they perform reliably in the real-world?

Software Engineering for AI

Developing systematic test and evaluation methodologies that ensure AI-enabled systems perform reliably in the real-world. My work spans test and evaluation of machine learning (ML) algorithms, deep neural network models (DNN), and large language models (LLMs), developing test methods and frameworks that address the unique challenges each technology presents.

• Test Generation: Test the implementations of traditional ML algorithms, pre-trained DNN models used in autonomous driving systems
• Test Set Reduction: Understanding the impact of sampling methods in generating test datasets
• Testing beyond Correctness (AI Assurance): Explainability, Fairness, Data Security
• Test Adequacy: Methods to assess data adequacy, test set adequacy
• Model Agnostic methods to perform test & evaluation (T&E) from development through deployment, including post-deployment T&E.
• Current Focus: Test & Evaluation of LLMs - Developing systematic evaluation methods, frameworks and tools for evaluating LLMs

AI for Software Engineering

With advances in LLMs, my research also explores the complementary direction: how LLMs can enhance and augment traditional software testing practices.

• LLM-Augmented Test Generation: Using LLMs for automated test case creation and input model generation
• LLM-Augmented Combinatorial Testing: Leveraging LLMs for efficient input parameter model construction
• Current Focus: Adapting off-the-shelf LLMs to generate input models

Explore My Work

→ Detailed Research Program & Contributions
→ Publications
→ CV

I am always eager to connect with colleagues, collaborators, and those interested in AI system assurance. Feel free to reach out for research discussions, collaboration opportunities, or questions about my work.