I'm the founding engineer at MakeTimeFlow, building the company's AI agent systems in close collaboration with the CEO: a voice-enabled coaching agent (Claude API + ElevenLabs real-time speech), the integration architecture connecting it to users' calendars and workflows, and the AI-native development workflows the engineering team runs on. I work across the full product lifecycle, from customer research and product strategy through architecture and deployment.
Before MakeTimeFlow, I led engineering on two competitively funded NSF research platforms ($1.2M+ combined) at the University of Houston. One supported emergency food distribution logistics for the Houston Food Bank; the other targeted counterfeit pharmaceutical vendor prediction. I owned the full AWS stack, microservices architecture, and data pipeline reliability. Before that, I built deep learning training pipelines at UChicago Radiology, where I led a small team developing U-Net segmentation models for detecting ablation zones during histotripsy treatment of renal tumors and improved baseline performance by 240%.
My path into engineering started in music. I studied audio design at Columbia College Chicago and wrote and produced music that's accumulated over two million views on YouTube. Studio work trains the same instincts engineering demands: signal versus noise, iteration under constraints, and knowing when something is done versus when you're just noodling. Those instincts carried over when I earned my M.A.S. in Computer Science at Illinois Institute of Technology.
Across all of this work, the pattern I keep seeing is that the model is rarely the hard part. The real challenges are in everything around it: data quality, pipeline reliability, evaluation, and the integration layer connecting predictions to the people who use them. That's where I spend most of my time, and it's the work I enjoy most.