About

My route into machine learning was indirect. I tried a few majors before landing on physics at UC Santa Barbara, and after graduating I took a job as a lab assistant at Pacific Diagnostic Laboratories. I processed specimens, verified orders, and tracked turnaround times in Epic. It sounds routine, but when something went wrong with the software, patients waited longer for results. That stuck with me, and it’s the reason I care about healthcare software that actually works.

I went back to school for an M.Eng. at UCLA, focusing on AI. There I joined Prof. Nader Sehatbakhsh’s Secure Systems and Architectures Lab and worked on real-time face anonymization for robots running on a Jetson Orin Nano. That project taught me that real systems have to respect memory limits, latency budgets, and privacy requirements all at the same time.

After UCLA, I volunteered in Prof. Wenbo Wu’s lab at Johns Hopkins, working on causal inference. I studied how the design of deep nuisance models changes treatment-effect estimates in double machine learning, which meant running and keeping track of 2,000+ model configurations on a SLURM cluster. In 2026 I also worked in Prof. Anind K. Dey’s lab at Georgia Tech on mental health assessment from wearable data, and built a self-hosted LLM system for Econ One Research as a consultant.

In August 2026 I started a PhD in Computer Science at UMBC, where I work with Prof. Dong Li in the FSI Lab on the security of foundation models for physiological signals like EEG, ECG, and wearable data. It pulls together the threads above: privacy, reliability, and models that clinicians and patients can depend on.

If you want to talk about any of this, email me at charleysanchez@umbc.edu.