I’m an AI researcher and engineer. I’m currently at OpenAI, building out the new monetization products (the ads in ChatGPT). I was previously at Google, working across a few different Ads products, most notably on applying LLMs to ads, and ads to LLMs.

My time these days is mostly allocated product-focused research, but I still work on ML theory. I’m particularly interested in quantifying and correcting the failure modes when applying ML to scientific problems, identifying when you can and can’t learn the underlying governing equations.

The pageload animation is the Stable Diffusion in-painting process. The first frame is the initial random image, and each frame is one step of denoising towards a top-of-fold screenshot. It’s just a pre-rendered video with two variations for mobile and desktop. Inspect it here.

The site icon is a heatmap of an ML weight matrix learning the 1D Poisson equation. You can see an explanation on my research page.

Some of my active work can be found on Github and in my hosted presentations. There are links to as many of my publications as possible on the CV page of this site. Conference presentations also have links to the recordings or slides. You can also find different things I work on around the internet:

google scholar Papers
afqueiruga LinkedIn
afqueiruga Research codes
goblinfightsraccoon Art
drgobble Keyboard builds