Hi π I'm Alex and this is my blog, which is mostly about biology and computers.
I completed my PhD in using ML for protein design in 2017 π and worked for 6 years at Ginkgo Bioworks π±, where I designed millions of proteins across hundreds of functional classes with AI π€, designed more synthetic DNA than anyone on Earth π§¬, and focused on protein therapeutics designed with generative ML βοΈ.
π» 2024. I'm currently open to opportunities contributing to a team working on ML for biology. Please get in touch by email (itβs on my GitHub profile) if youβd like to chat!
π³ 2021β2023. As Sr. Engineer for protein ML at Ginkgo, over two years I led dozens of protein ML projects and was responsible for benchmarking, scaling, and development of generative ML models for protein design and discovery. My main responsibilities included designing lots of proteins, implementing and onboarding new methods from the literature (diffusion models), scaling existing approaches (sequence and structure transformers), and detailed benchmarking of the strengths and weaknesses of existing methods for specific problems to develop βevalsβ for biology. I also solved some fun biology problems, like combining molecular dynamics and deep learning to design multi-state enzyme reactions.
π² 2017β2021. I joined Ginkgo Bioworks as a Protein Engineer focused on enzyme design and discovery using machine learning approaches. I designed hundreds of large learning libraries of protein therapeutics in iterative cycles employing ML with an extremely high success rate. In addition to designing millions of proteins across hundreds of functional classes, I also created a machine learning enzyme engineering data platform using containerized, GPU workflows to enable benchmarking and automated iterative design which has now been used across thousands of iterative ML projects.
π± 2013β2017. My PhD was focused on advancing the field of machine learning for protein design via leading a team to the construction of a large, high-quality benchmark dataset that enabled the improvement of ML protein design algorithms. My adviser was Justin Siegel at the University of California, Davis.
π 2010β2013. I made wine in Napa Valley (3 harvests) and decided to go to grad school
π§ͺ 2006β2010. BA at Bard College in biology, for two years I was a teaching assistant for organic chemistry (which remains my favorite subject)
Papers, patents, and other publications:
- Writing on biology and computers, this site, where I write about biology and computers
- Google Scholar lists my scientific publications on ML for protein design, and Google Patents lists many of my inventions related to designed enzymes achieved over my career thus far with many brilliant coworkers
- I do lots of side projects for fun using ML, with a particular focus on things that are educationally useful, fun, and hackable