I Was There When: AI helped create a vaccine

The entire process from start to finish can cost billions of dollars and take up to a decade to complete. It fails in a lot of cases. There are many diseases out there that don't have a vaccine or treatment for them. It's not like people haven't tried, it's just that they are.

How can we shorten those timelines? What can we do to target more things? That's how I got into the company. My background is in data science. I hold a PhD in information physics, which is related to data science.

I started working for the company when it was a small one. We were trying to figure out how to target a bunch of different ideas at once, run some experiments, learn really fast and do it again. Let's run a hundred experiments at the same time so that we can learn quickly and move on to the next stage.

If you want to run a lot of experiments, you have to have a lot of messenger RNA. We needed to integrate all of the things we built out. We need systems to drive all of those. As you capture data in these systems, that's where you start to see the emergence of artificial intelligence. Let's use the data from the experiment to make some predictions instead of just capturing it.

Scientists don't want to look at data over and over again so let's take out decision making from them. Let's use their knowledge. It would be great if we could automate their analyses and do a better job of predicting outcomes and improving the quality of our data.

When Covid showed up, it was a powerful moment for us to take everything we had built and everything we had learned, and apply it to this important situation. When this sequence was first released by Chinese authorities, it only took 42 days for us to identify, you know, these are the kinds of changes we want to make. We want to target this molecule.

It takes 44 days from that point to build up a human safe manufacturing process and then send it off to the clinic. A lot of people were surprised by how quickly it moved. We built this engine for 10 years to speed up research. It didn't end there.

How can we use data science and artificial intelligence to inform the best way to conduct our studies? In order to prove that this works, we had to do a large phase three trial of 30,000 people.