Hello, I’m Annmaria! I build learning systems that can be brought to the physical world. Previously, I worked on ML for VLSI at NVIDIA, hardware performance analysis at Intel, and small-molecule drug discovery at Integrated Biosciences. I also developed post-training methods for adapting language models for kernel generation under Azalia Mirhoseini at Stanford’s Scaling Intelligence Lab. My work was supported by AMD.

I believe hardware innovation sits upstream of progress on some of the world’s most important problems (human health, national security, etc.). But hardware innovation can only move at the speed of manufacturing, and American manufacturing runs on iteration cycles measured in months, using processes that haven’t meaningfully evolved in decades. My vision for the future includes a world where manufacturing a complex, high-precision part is as fast as compiling software.

I grew up in the Bay Area and attended The Harker School. In 2023, I was recognized as the top female graduate in California and named a U.S. Presidential Scholar by President Joe Biden. At Stanford, I studied CS/EE and was advised by Dawson Engler.