Agile Research Group

The Agile Research Group advances software engineering for machine learning and cybersecurity by building reliable, explainable and secure AI systems. Its research spans requirements engineering for ML (elicitation, verification/validation and traceability), formal methods for models, synthetic datasets, ontology/knowledge graph-driven data engineering, quantitative explainability for vision and speech, and LLM/RAG safety (attack surfaces, defenses and provenance). Applications include aerospace and safety-critical domains, with outputs that range from benchmarked datasets and tools to peer-reviewed publications and industry-transferable methods.

Lab Goals

• Advance trustworthy AI engineering by developing methods for requirements engineering, verification/validation, dataset provenance and secure deployment of ML systems.
• Bridge AI and cybersecurity by investigating vulnerabilities in modern systems, creating defenses and applying formal methods to critical domains such as aerospace.
• Empower students at all levels, from undergraduate through Ph.D., to contribute to cutting-edge research, gain publication experience and build skills in software engineering for machine learning.

Capacity

Researchers execute end-to-end ML engineering studies — from requirements and dataset design to formal verification, secure LLM/RAG evaluation and deployment-grade tooling — supporting multiple concurrent projects and delivering publishable results, reproducible pipelines and validated datasets/models.

Equipment

• High-performance GPU Workstations (4 total): Each equipped with modern NVIDIA GPUs for deep learning model training, dataset generation and large-scale experiments in machine learning and cybersecurity.
• Quadcopter Drones with Cameras (2 total): Each drone operates with GPS, video and sounds capture to support experiments in cybersecurity and machine learning for uncrewed aerial systems.
• HackRF: Hardware for software-defined radio to support cybersecurity experiments on the drones.