Lawrence Zou

About

“Life is really simple, but we insist on making it complicated.” — Confucius 🎶 Currently building a machine learning jazz improvisation generator! Built the model from the ground up in PyTorch, implementing a small ~3-5 million parameter model using the transformer architecture, and trained on 3 GB of MIDI music data. I'm someone who appreciates how technology can elevate our experiences. Evolving technologies will simplify our quest for information, allowing us to better formulate our understanding. I view machine learning as one step in bringing more information into our grasp, but ultimately, by the way current representations of data (text, images, videos) are encoded, as well as choice of ML architecture, there are limitations as to how much information can be encoded and manipulated. Interested and passionate about exploring the boundaries of capabilities of current ML systems in data encoding and architectural choices, as well as working to discover new ways of supplying quality data for training. I strive to innovate from fundamental ideas first. Once strong goals are established, I will look to find effective solutions that brings results quickly, allowing for a faster rat

Experience

Education