Daphne Koller didn’t anticipate becoming a drug hunter when she joined Stanford University as a professor of machine learning in 1995. But after stumbling into biology a few years later because of the novelty of the small but diverse data sets it offered, the field quickly drew her in. By 2016, she was the chief computing officer at Alphabet’s ageing-focused biotech Calico. She then founded the biotech insitro in 2018 to explicitly generate high-quality biological data sets at high throughput to enable the application of machine learning to drug discovery. She spoke with Asher Mullard about her vision for machine learning in drug discovery, the need to view data as a core asset of a business rather than a byproduct and how to foster computational excellence in biopharma.