Managing AI and data science: Practical lessons from big pharma
A top AI leader from Novartis shares his advice on how to lead an AI team. He discusses AI in drug discovery, but the challenges and solutions are applicable to every business leader. Watch the video and read the transcript to learn what you need to know!
Data science and artificial intelligence are adding a new dimension to drug discovery and development, emphasizing computation and machine learning. Given this shift, pharmaceutical companies are actively building infrastructure, data, tools, and teams to bring together data scientists with biology and life science experts.
Pharma and biotech innovation offer a glimpse into how large organizations integrate AI tools and techniques with traditional subject matter experts who possess a deep understanding of the underlying problems to be solved.
To gain an insider's perspective on how pharma companies use AI and machine learning, I invited Dr. Bülent Kızıltan to join episode #717 of the CXOTalk series of conversations with people shaping our world. He is Head of Causal & Predictive Analytics, Data Science & AI, at the Novartis AI Innovation Center.
Dr. Kızıltan is one of the most articulate people I know on managing and leading AI efforts, so watch the video for helpful and practical advice on managing data science and AI teams.
Drug discovery and development have slowed down over the last five to ten years because of high costs and because scaling up is very difficult. We hope AI can come to the rescue, so many pharmaceutical companies are investing in this area.
AI and data science, in general, can operate in one of two ways. One way is to be use-case driven, providing those services to business units.
The other case is where we position ourselves at the intersection of academia and the business units. Academia is creating the knowledge, technological development, and infrastructure to scale things up.
In data science, typically people believe that teams work with big data and when there is a limited amount of data, the value proposition declines. [However,] we cover the whole spectrum, from small data to big data because those terms are vaguely defined, and we don't have a clear way to quantify small and big data.
We've built core capabilities to extract forecasting information from limited data all the way to big data, as we call it. We extract information from limited data in the domains of healthcare, biotech, and medicine.
Talent management and diverse teams in pharma
AI innovation, especially data science, is a very inter-disciplinary and multi-disciplinary domain. We want to attract talent from different disciplines who can bring the mindset of their own domain into our operations.
Certainly, data science and machine learning core capabilities are necessary, but we're open to all backgrounds. As you might know, I was trained as an astrophysicist and have studied neutron stars and blackhole astrophysics for most of my career. But in that domain, I worked very closely with applied mathematicians, machine learning pioneers to bring those technologies into the domain of astrophysics.
We are very cognizant that diversity is necessary to think out of the box and innovate in AI. Currently, we're growing our teams and looking for talent to bring in core capabilities that are necessary with machine learning, but they can come from physics, mathematics, psychology. I had worked with people coming from sociology, economics -- you name it.
Recently, I was part of a global benchmark study looking at companies across the globe, thousands of companies from different domains. We've seen that culture and leadership are critically important in success.
If you want to sustain the value proposition for the long-term, you have to build culture and company around that. This is necessary to make an impact on the ground, to reimagine medicine.
Using AI and data science in drug discovery and precision medicine
Customizing medicine and treatment is a problem faced by the larger biotech and healthcare industry. We believe that customization, at scale, can only happen with the help of AI.
Precision medicine is one focus area for every pharmaceutical company, biotech, and healthcare company. Leveraging the technology that is developed in the domain of AI is critically important. It will redefine that whole domain.
Developing and discovering compounds and drugs (generative chemistry, to be more technical) is an area where AI and machine learning are making an impact. There are many companies using data science, AI, and machine learning to augment the development process and discover new compounds.
Previously those were done only in the lab. It was a painstaking and difficult process. With AI, we may be able to do that all in-silico, on the computer. We can simulate. We can come up with a priority list of compounds and then talk to our domain experts about whether what we find makes sense or it's totally crazy.
Managing bias in data science and AI
We engage domain experts from different disciplines to address this problem: there may be sampling biases, algorithmic biases, or data-driven biases.
We need to address all those things early on. Once we come up with a forecast or prediction, we take specific steps to ensure that we are not biased to a level that affects the decision-making.
It's an actively developing domain, and I have yet to see a strong, quantitative perspective and methodology that will help us address it. [Today, we address bias] on a use case-by-use case basis. It's a great question.
CXOTalk presents in-depth conversations with people shaping our world. Thank you to my senior researcher, Sumeye Dalkilinc, for assistance with this post.