Coworking: Anna Marie Wagner believes AI will excel in biology

The Head of AI at Ginkgo Bioworks says “DNA is just another language.”
article cover

Anna Marie Wagner

· 5 min read

Coworking is a weekly segment where we spotlight Tech Brew readers who work with emerging technologies. Click here if you’d like a chance to be featured.

How would you describe your job to someone who doesn’t work in tech?

As is common in hyperscaling companies, I’ve had a different job just about every year I’ve been at Ginkgo. Ginkgo is building the leading horizontal platform for biological engineering and biosecurity, and although we’re in the life sciences, we have a tech platform business model. Think about it like programming, but instead of computers, we’re programming cells (which almost magically can read and write digital code in the form of DNA, which is made up of As, Cs, Ts, and Gs instead of zeroes and ones).

Coming out of a decade in private equity, where I was investing in tech companies, I was hired to be the person who would get big, transformative acquisitions and partnerships done for the company. But in addition to that job (and we’ve done more than a few acquisitions!), I also stepped in as interim CFO for just under a year, took the company public, and launched our AI efforts with a Google partnership. Today I still lead corporate development (M&A, big strategic partnerships, capital markets, etc.) and am also Ginkgo’s Head of AI, where I focus on helping our customers benefit from the rapid advances in AI. DNA is just another language, but it’s one that humans didn’t invent, and so in many ways, AI can advance our capabilities and understanding much faster in biology than it can in human languages. There’s so much more to come, and it’s a wildly exciting time to be in this space.

What’s the most compelling tech project you’ve worked on, and why?

Ginkgo is in the business of taking a customer’s important, but seemingly impossible, projects and making them possible. I’m constantly blown away by some of the technical breakthroughs our team regularly makes in diverse areas.

One of the projects I had a more direct hand in was establishing Ginkgo’s foundation for AI. In the course of our work doing cell programming year-in and year-out, Ginkgo has been amassing a very large database of metagenomic data (think: genomes from lots of different microbes, which code for billions of different proteins) as well as large structured datasets, labeled with experimental results. As new LLM architectures make it relatively easy to integrate vast amounts of data and yield emergent properties of interest, we developed a perspective that ultimately the algorithms are likely to be largely open-sourced commoditized, and that data is what is really what matters right now for AI in biology. That is, having the data to train and fine-tune models for practical applications is going to be the differentiating factor for companies looking to make real-world applications possible. The good news: Ginkgo has the data, and we have the capability to generate even more data in a cost-advantaged way. So we decided to expand our AI strategy and established a really interesting partnership with Google to build foundation models for biology. In doing so, we’ve basically taken compute costs out of the equation when it comes to training these large models, and I’m really excited about what the combination of Ginkgo’s data and a large compute capacity can bring to bear.

Keep up with the innovative tech transforming business

Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.

What technologies are you most optimistic about? Least? And why?

You know, it sounds really boring, but I’m most excited about advanced laboratory automation. The life sciences industry has been held back by the fact that humans are reasonably slow, generally have at most two arms, and make relatively frequent and stochastic errors. Ginkgo has been investing for 15 years in the physical infrastructure and software necessary to automate massive biological workflows, allowing us to generate consistent data across many experiments. This is the foundation for our AI work and is what, I believe, gives us a strong advantage as we start adopting those technologies.

What’s the best tech-related media you read/watch/listen to?

I’m a big fan of the Acquired podcast and the 7 Powers framework they apply at the end of each episode, based on the book by Hamilton Helmer. 7 Powers is a terrific way to honestly assess the sources of long-term competitive advantage in a business—we use it regularly at Ginkgo, and I recommend it to others.

What’s something about you we can’t guess from your LinkedIn profile?

I’m an excellent cook and baker, and once spent six months of consistent practice mastering the French macaron (including making scatter plots of different recipes and the ratios of the key ingredients therein—baking is basically just chemistry, sometimes with a dash of biology thrown in). I may have some essence of Elizabeth Zott in that domain. (I also make a mean lasagna.)

What do you think about when you’re not thinking about tech?

Well, in addition to the meaning of life and future of humanity, I’m probably either playing holiday songs on the piano during inappropriate seasons (which my husband would argue is anytime before Thanksgiving, whoops!) or figuring out when I can sneak in a bike ride.

Keep up with the innovative tech transforming business

Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.