Muhammad Tehmasib Ali Tashfeen

I am a PhD candidate in the School of Computing at Wichita State University, working in the Software Analysis and Intelligence Laboratory under the supervision of Dr. Zhiyong Shan. My dissertation, “Automated Detection of Android Lifecycle Anti-Patterns: From Static Analysis to Fine-Tuned Large Language Models,” develops principled techniques for finding reliability and security defects in mobile software by combining static program analysis with domain-adapted code language models.

I designed and built LeakScope, a Soot-based static analysis framework that translates Android lifecycle obligations into detection rules for memory-retention and resource-management anti-patterns. In our large-scale empirical study, LeakScope analyzed 627 open-source Android applications and flagged 485, with app-level detection coverage compared against Android Lint, SpotBugs, and Infer. In a parallel line of work, I constructed 3,152 prompt-response examples and developed LoRA fine-tuning pipelines for CodeLlama, Llama-3, and StarCoder2, evaluated against GPT-4o, to investigate whether open-source LLMs can reliably identify lifecycle violations from source code.

Prior to my doctoral studies, I spent three years in industry cybersecurity, including roles as a Senior Consultant (Technology Risk) at Ernst & Young Ford Rhodes and an Information Systems Auditor at The Bank of Punjab, where I evaluated security controls across 100+ enterprise systems and led audits for major telecommunications and financial institutions. I hold a BS in Electrical Engineering from FAST-NUCES, Peshawar.

Research Interests

• AI-assisted software engineering
• Static program analysis
• Android application reliability and security
• Domain adaptation of code language models
• Empirical software engineering and reproducible evaluation