Passion fuels every line of code and every inquiry in my research—driven by a relentless pursuit to redefine how intelligent systems can enhance reliability, testability, and developer trust in complex software ecosystems. My work bridges AI, software engineering, and immersive technology, with a focus on real-world applicability in environments like virtual reality, edge systems, and intelligent automation.
At the University of Michigan-Dearborn, I led the development of VRTestSniffer and CPatMiner, two advanced Java-based tools that analyze Unity-based VR projects for test smells, refactoring opportunities, and underlying code issues. These tools apply static and semantic analysis to detect critical quality flaws such as:
Obsolete test logic
Unrefactored assertion blocks
Complex conditionals
Duplicated test scaffolding
Improper coupling between test and scene logic
With CPatMiner, we went beyond detection and focused on actionable bug patterns, such as:
Hidden scene state leaks caused by implicit fixture sharing
Assert logic tied to outdated object references
Fragile object instantiation across test setups
Repetitive camera-state initialization blocks
These insights helped build a foundation for recommending automated refactoring—which I’m actively expanding in our ongoing VR Refactoring Research, aiming to develop an AI-assisted code transformation engine for Unity and C# test scripts. The goal is to automatically refactor bad test patterns into reusable, maintainable structures optimized for immersive platforms.
This research, submitted to ASE 2025 and shared with the U.S. Army simulation testing teams, demonstrates how static smell detection, LLM-assisted tracing, and targeted refactoring can streamline QA pipelines in high-stakes virtual environments.
Complementing this work, my earlier research into edge-based content delivery systems explored how ML inference can be optimized at the router level to personalize recommendation systems while minimizing latency. This project shaped my broader interest in distributed AI systems and explainable automation.
For me, research isn’t just about innovation—it’s about building systems that are interpretable, maintainable, and impactful. Whether it’s helping developers identify bugs faster, ensuring military simulation stability, or crafting refactoring logic to prevent regressions, my work is rooted in the belief that AI and engineering should amplify human clarity and trust—not replace it.
My Research page: Research pageÂ