About
I’ve always been drawn to the space where art and machine learning meet. Now I’m building Kivo—a multimodal agent that understands graphic tools and learns to design through action.
Can a model learn design inside the same tools—and under the same constraints— as a person?
Since 2020, I’ve worked across generative modeling—from normalizing flows for diverse super-resolution to transformer-based image generation. I was a first author of Styleformer—the first style-based Transformer GAN for image generation—built as a curiosity-led side project without a professor or a lab agenda. The work introduced the first use of Linformer for image generation and was accepted to CVPR 2022.
In 2023, I founded OptimizerAI to turn original generative-model research into creative systems. I built a flow-matching sound foundation model from scratch out of curiosity; that research became the company’s foundation. OptimizerAI was backed by a16z Speedrun. Today, OptimizerAI is where I’m building Kivo and exploring what happens when models work inside the same tools people use.
Now I’m building graphic environments where agents can see the canvas, use tools, and receive feedback. I’m interested in whether reinforcement learning can teach a model not just to produce an image, but to make design decisions more like a person.
Current thread: graphic environments × reinforcement learning.