An essential part of our philosophy for success in drug discovery is our quarterly hackathon, a hands-on technical challenge that our entire team works together to solve. As a data and science-led organization, these hackathons keep our team working collaboratively across departments and integrating innovations from across the fields of bioscience, pharmaceuticals, computing, and quantum technology.
For example, the goal of a recent hackathon was to create new AI/ML components for our
drug development pipeline. This process allowed the team members from chemistry, computer science, and dev-ops to share their expertise in theoretical and technical matters. With these complementary skill sets, we learned about cutting-edge work outside of our day-to-day tasks and developed new tools that combine these technologies with the power of quantum computing to accelerate the finding of promising small molecules for customers and collaborators.
Our best practices for hackathons include the following recommendations:
Prior to the start of the hackathon, we make sure that all team members have compatible computers and software. To accomplish this we make a docker container for each hackathon that contains the necessary elements and tools to attempt to solve the problem. By putting all of the tools needed to engage the problem into one place, we avoid the time-consuming issues of installation and software compatibility across many machines, situating all team members at the same starting point at the start of the first day. At this point, our tradition is to celebrate the start of the hackathon by opening our boxes of hackathon-themed candy, delivered to us from our home office in Durham, NC.
The goal of the hackathon should be a hands-on experience that is relevant to all team members and has an experimental phase for testing. We start the day with a brief overview of the goal of the hackathon, divide the team into smaller groups that include people from both chemistry and dev-ops, and then go straight to the code to start hacking. Since we give ourselves only two days to make a beta version of a model, we work together in shared environments and keep a rolling discussion of ideas and tests of those ideas. Every two hours, all of the smaller teams meet back together to say what they have tried, what they found, and what they will do next. This approach keeps us coordinated while still having smaller groups work on parts of the overarching problem. At the end of the hackathon, each group presents its model and the results of testing.
Hackathons are immensely valuable to our process because they allow us to make major advances to our system in a short amount of time. While it is a major ask to call every person away from their usual work for two days, we have always made tremendous progress that keeps our technology moving forward and our team energized and inspired. One of our chemistry team members, Anna Petroff, commented “I’m always amazed at how much progress we can make on a project during a hackathon. When we can take some time to work across groups, we can go from an idea for a feature to a working beta version in 48 hours.” This progress, combined with the comradery developed from trying to achieve a challenging goal as a team in a short time frame, makes hack-a-thons an invaluable tool for us at POLARISqb.