Andrew D'Angelo

• Developed and enabled system-level software for AMD's Helios platform, enhancing next-generation datacenter solutions. • Led platform bring-up and feature enablement, ensuring seamless integration of hardware and software components. • Optimized performance and debugged issues across hardware and software boundaries, contributing to overall system efficiency.

As a Senior Software Program Manager, I lead and actively contribute to the development of our real-time Factory Floor Dashboard (FFD), a critical platform used by global manufacturing teams for SLT (System-Level Test) and RLT (Rack-Level Test). I own the end-to-end delivery of features, from architecture and requirements through development, implementation, testing, and deployment, while working directly in the codebase to unblock teams, improve reliability, and accelerate releases. In parallel, I design, build, and scale our Grafana dashboards that aggregate server logs, test results, and telemetry from multiple factory sites in real time. These dashboards power operational monitoring, anomaly detection, and performance analysis. My work spans backend services, data pipelines, and observability tooling, with a strong focus on scalability, maintainability, and production-grade reliability, while collaborating closely with engineers and operations to ship high-impact solutions.

I worked as a researcher with the AFRL for two consecutive summers. During my first summer with the Air Force, I researched semantic-based security solutions for smart home and smart building environments. I leveraged semantic information from API documentation, open-source code, and user activities to generate correlations between smart devices in IoT environments. This research was published in my paper, "Discovering Complex Correlations among Multiple IoT Devices in Smart Environments," presented at IEEE Globecomm. In the summer of 2023, I advanced this research by designing a novel transformer-based correlation miner using attention weights from a fine-tuned BERT transformer. I simulated real-world attacks and achieved a detection accuracy of 96.59% and recall of 97.38%. These results were published in "Transformer-based Compound Correlation Miner for Smart Home Anomaly Detection" at the IEEE CloudNET 2023 conference. Additionally, I developed a GRU-autoencoder attack detection system for IoT smart building systems, which attained 97.97% precision, 98.18% accuracy, and a 0.05% false alarm rate.

I co-founded SecureMeeting, a 501(c)(3) nonprofit, with the mission to provide free and secure video communication for people confined to their homes during the COVID-19 pandemic. Our goal was to ensure that everyone could connect confidently, without compromising their privacy or personal data. Since its launch, SecureMeeting has been used in over 65 countries, with an average of nearly 500 daily users. I led the engineering of a decentralized microservice architecture with custom APIs for video room management and additional features. We implemented Agile methodologies using JIRA, managing 2-week sprints with an average of 30 story points per sprint. Our efforts created a lasting, secure, and accessible communication platform for users worldwide.