ml researcher & engineer
I'm building umph.ai, an AI-powered operating system for the aesthetic industry. Think of it as the intelligence layer that runs the whole thing.
In my free time, I train models and run experiments on ideas that probably won't work. Until, occasionally, they do.
Before this, I did my PhD in CS focused on data privacy, specifically using VAE-based generative architectures to make ML models fairer and less biased. The technical framing was privacy; the deeper motivation was fairness.
After that, I worked as an AI/ML engineer at Oxy. Good work, sharp team. But I wanted to build something of my own, so I left.
What I keep coming back to: intelligence and generalization — what it means for a model to actually understand something, and how far today's systems are from that.