Research & Development Projects

An end-to-end Python code-generation model fine-tuned from Qwen2.5-Coder-1.5B across three stages. DoRA SFT on a curated CPython/Flask/Requests corpus reached 70.7% HumanEval (+8.2 over base on spec-to-code), a GRPO (RLVR) pass with an execution-grounded reward lifted MBPP to 72.5%, and a hybrid RAG layer (Qwen3 embeddings + LanceDB dense/BM25) injects exact API signatures to fix the model's weakest axis.

A threat taxonomy and architectural attack model for agentic AI coding tools (Claude Code, Cursor, Copilot Workspace) and the Model Context Protocol (MCP) ecosystem. Defines seven threat classes unified by a "trust laundering" meta-concept, formalizes novel primitives including a JSON Encoding Layer Bypass (JELB) and an Agentic CLI Relay, and maps 17 disclosed CVEs onto the taxonomy with STRIDE analysis and a layered defense framework.

Designed a novel loss formulation mitigating catastrophic forgetting in continual learning. Engineered a phase-aware loss strictness scheduler driven by expected output entropy, combined with a Leitner-box retry-escalation mechanism. Evaluated across TAMU HPRC clusters using synthetic datasets and GLUE benchmarks on BERT/GPT-2.

Architected secure middleware for Model Context Protocol (MCP) servers and clients. Implemented a post-quantum Kyber-768/X25519 hybrid Key Encapsulation Mechanism (KEM) integrated into a Noise XX handshake. Built decentralized identifier (DID) identity verification, a 3-tier inputs sanitizer (homoglyphs, mixed-script, control characters), rate limiting, and semantic similarity guardrails to prevent injection, spoofing, and MITM attacks.

Conducted an enterprise threat modeling study mapping risks of unauthorized LLM use (Shadow AI) to industry frameworks (NIST CSF 2.0, NIST AI RMF, ISO/IEC 42001, OWASP Top 10 for LLMs). Defined multi-tiered mitigation controls including tenant restrictions, Semantic Prompt Firewalls, and sandboxed local RAG systems.

Fine-tuned and evaluated self-correction behaviors in LLMs. Executed LoRA SFT on Qwen2.5-Coder-7B-Instruct using CodeContests. Analyzed severe out-of-distribution performance shifts on HumanEval (57.93% vs 50.61% baseline), MBPP (4% vs 0%), and Codeforces (8.7% vs 0%), documenting the fragility of SFT correction patterns.

Built a hybrid dependency upgrade analysis engine in Go combining native AST structural call-graph extraction with LLM reasoning (Gemini/Gemma) to evaluate upgrade safety. Natively parses Go, Python, and Node.js codebases, filtering cosmetic diffs while flagging structural breaks, signature alterations, exception semantics, and call-chain removals. Features CI/CD integrations for GitHub Action annotations and Slack Block Kit webhooks.

Developed a secure browser extension running DistilBERT locally via ONNX Runtime Web. Engineered a Hybrid Scoring and Visibility Engine to detect prompt injections and sensitive PII leaks, defeating tag-splitting, indirect injection, and asynchronous exfiltration channels.