research
Grants & Funding
Enhancing Career-Ready Skills in Cybersecurity Through Integrated ELK Stack and Machine Learning Labs
ECU 2026/2027 Teaching Grant · PI · Summer Stipend · 2026
Open-Source Intelligence Sandbox (OSINT)
Civil-Military Innovation Institute, Inc. · PI · $261,847 · 2024
Misbehavior Detection in Vehicular Communication Networks
Toyota North America Inc. (sub-contracted through UNL) · PI · $20,000 · 2020 - 2021
Accelerating Credentials of Purpose and Value Grant Program
Texas Higher Education Coordinating Board (THECB) · Co-PI · $410,000 · 2022
GET PHIT: Gaining Equity in Training for Public Health Informatics and Technology
Office of the National Coordinator for Health Information Technology · Co-PI · $265,000 · 2021
RSCA Grant
University of Texas Permian Basin · PI · $3,000 · 2021
UT STARs Grant
University of Texas System · PI · $80,150 · 2020
Current Projects
Trustworthy Misbehavior Detection for Connected and Automated Vehicles
Connected and Automated Vehicles depend on trustworthy V2X communication to detect malicious safety messages and enable safer, coordinated driving. Most AI-based Misbehavior Detection Systems (MDS) remain black-box, unrealistic in their assumptions, and hard to deploy in real time. This project builds a unified MDS framework combining low-latency detection, robust model updating, and interpretable response for real-world CAV deployment.
Multi-Agent LLM Framework for Trustworthy IoT Threat Intelligence
Traditional LLM-based question-answering systems often hallucinate or return outdated answers, a serious risk in IoT security where vulnerabilities span heterogeneous devices and protocols. This project develops a multi-agent framework that decomposes security queries, retrieves evidence from structured sources like MITRE ATT&CK for ICS, and coordinates reasoning agents to synthesize grounded answers.
Prompt Injection Guardrails for Home IoT Assistants
LLM-powered smart home assistants are vulnerable to prompt injection and role confusion from untrusted inputs. This project proposes a layered architecture that incrementally adds authentication, policy enforcement, sanitization, and post-LLM gating, substantially reducing risk even when the model itself produces unsafe commands.
Benchmarking Network Education Platforms
Networking education requires students to build both conceptual understanding and hands-on diagnostic skill, yet meaningful lab access is difficult to sustain across simulation, emulation, remote, and physical hardware environments. This applied study implements a standardized suite of CCNA-aligned labs across Cisco Packet Tracer, Cisco Modeling Labs, remote physical equipment (NDG NETLAB+), and local physical hardware.