A: Despite the marvels and triumphs of modern medicine many diseases remain untreatable. And the development of novel therapeutics to address these conditions isn’t getting any easier or cheaper, with clinical trial success rates hovering in the single-digits and the capitalized R&D spend per approved drug estimated at $2.5B. At insitro we are setting out to rethink drug discovery and development, leveraging recent advances across computer science, molecular biology, and automation to develop better and cheaper therapeutics in less time.

A: Insitro is a therapeutics company – ultimately if successful we will develop medicines that enable patients to live longer and more fulfilling lives. We are unique in two main respects. First, we are building an interdisciplinary organization from the ground up, based on the fundamental principle that by creating an integrated team of biologists, chemists, technologists, and computational scientists who speak each other’s language we can do far more than any group can do on their own. Second, we are investing heavily in data generation and the ability to generate data – effectively building a factory with the express purpose of producing data on which to do machine learning. Machine learning models are only as powerful as the data they are trained on – by generating datasets explicitly designed for machine learning and using the subsequent analysis to drive the next round of data generation we are closing the loop between in silico and in vitro methods.

A: I am the Founder and CEO of insitro. So much excites me about my work. I get to build a forward looking organization hiring and working beside incredibly talented people while delving into the complexities of biology, helping solve challenging machine learning problems, and designing novel systems to automate and standardize state-of-the-art biological experimentation. And in the end we are doing all of this in order to help patients and make the world a better place. What’s not exciting about that?

A: I believe we are at the beginning of a new scientific epoch resulting from the intertwining of data science and biology. This is driven by a range of advances across molecular biology and technology, including gene-editing, single-cell sequencing, automation, microfluidics, and advanced microscopy, paired with the enormous recent advances in machine learning and high-performance computing. Machine learning is solving a range of problems that I, as a machine learning researcher for 25 years, didn’t think would happen in my lifetime including human-level image captioning and language translation. While the advances across these two disciplines are each powerful on their own, it’s their synthesis that will allow us to gain new insights into underlying disease processes and ultimately create novel medicines that benefit patients.

A: First, we face the challenge of being on the cutting edge of science – human biology is incredibly complicated. Building science-based companies is exciting but requires bringing a lot of pieces together. We are pushing boundaries in each of several disciplines and it’s a challenge to walk the line of innovation while also making sure we can scale what we’re doing. We’re also facing the challenge of finding and hiring the right people. There are a limited – but luckily growing – number of people that speak both machine learning and biology. Identifying and attracting them is critical to the success of any organization looking to operate at the intersection of the two. Finally, it is key to build the right culture: one where people trained in very different languages work effectively together as a single team, so that the whole of the organization is considerably bigger than the sum of its part.

A: This is an incredibly exciting time to be at the interface of biology and machine learning – we are really just at the beginning.