Afsah Anwar

Afsah Anwar is a tenure-track Assistant Professor in the Department of Computer Science at the University of New Mexico (UNM), Albuquerque, leading the Beyond Defense Lab focused on Internet security, malicious activity analysis, and incident response frameworks. Her work bridges technical cybersecurity research with societal impact assessment.

Her research spans Malware Analysis, Threat Intelligence, and Vulnerability Management, with specialized expertise in Machine Learning-based defenses for IoT/mobile platforms, Internet censorship measurement, space security, and web fraud detection. The lab emphasizes real-world applicability through tools that address concept drift in malware detection and robustness against adversarial attacks.

Analysis of her 2020-2025 publications reveals dominant themes: IoT malware robustness under adversarial conditions (40% of output), Internet measurement studies including censorship and routing security (30%), and vulnerability database refinement (15%). Key methodologies integrate graph-based analysis, time-series measurement, and multimodal LLMs for interpretable security reasoning.

Scientific Awards:

• Best Paper Award at WISE 2024

Research Grants:

• NSF SaTC Medium grant (August 2025) on VPN security with Roya Ensafi (UMich) and Jed Crandall (ASU)

Advising:

• PhD Students: Enrique Sobrados (Mobile Security), Ramsha Anwar (Mobile Security/Privacy), Dheeman Saha (Censorship/Internet Measurement; co-advised)
• Master's: Jack Vanlyssel (NSF SFS Scholar, Space Security), Muhammad Danish (Web Security; CRA Outstanding Undergraduate)
• Undergraduate: Muneeb Ahmad (AI/Cybersecurity)
• Mentorship: Weekly CTF events, Sandia National Labs collaborations (Ross, Doakes, Hassan), and dedicated support for students from Bihar, India

The Beyond Defense Lab operates from Farris Engineering Center #2150, hosting industry experts (e.g., Meta's Dr. Abusnaina, X's Dr. Saad) and developing healthcare security frameworks with Dr. Sazzadur Rahman. Current projects include Russian censorship analysis using time-series methods and multimodal LLMs for vulnerability reasoning (LLaVul).