Chief Science Officer Ivan Drokin, who has worked in various fields of AI, from finance to biotech, for over 15 years, gives advice on how to get into the industry and what it takes to manage research and development (R&D) teams.

Over the course of his career, Mr. Drokin's team of researchers developed and implemented over 50 complex projects, ranging from lung cancer screening to clinical trial data analysis. Mr. Drokin also created and patented several solutions in the field of biotechnology.

One of his notable accomplishments was implementing retrospective lung cancer screenings on CT scans collected during the COVID-19 pandemic, identifying hundreds of new suspicious cases. Mr. Drokin's team dedicated a cluster of patents to modeling and processing patient data, particularly in medical imaging, which became the cornerstone of the company's product portfolio.

Mr. Drokin has also undertaken cutting-edge projects for international industry leaders, including LG and Arrival. 

Mr. Drokin is recognized by the scientific community through his published scientific papers recognized with Best Paper Awards at international conferences AIST.

Ivan Drokin
(Photo : Ivan Drokin)

For those aspiring to pursue AI research, Ivan Drokin stresses the importance of a robust mathematical foundation, including calculus, probability theory, and optimization. He highlights the significance of academic background, university selection, conference participation, and publication venues for engaging in research.

Mr. Drokin points out leading institutions such as Stanford University, UC Berkeley, California Institute of Technology, the University of Toronto, Imperial College London, and prominent AI labs within companies like FAIR (AI lab at Meta), Google, DeepMind, OpenAI, and Amazon, as hubs of excellence in the field.

For those who would like to enter the industry, Mr. Drokin notes the increasing accessibility of roles related to AI. He suggests that understanding solution assembly from different components facilitates industry entry, making access more democratic.

Mr. Drokin also advises staying updated on industry trends in AI and ML and observing expert methodologies. He recommends seeking guidance from experienced practitioners as their insights can offer valuable directions. Working in R&D teams involves unique challenges, for example, the outcomes often remain unpredictable and require an adaptive approach to project management, such as decomposing projects into testable hypotheses.

Long and complex projects that face setbacks can greatly demotivate teams, which forms another challenge that requires addressing. 

Therefore, effective communication among team members, support mechanisms, and a culture that embraces failure as integral to the research process remains crucial. This management style significantly differs from, for example, software development, according to Mr. Drokin.

Regarding the evolving AI landscape, Mr. Drokin observes a shift towards development-centric approaches driven by large generative models. 

With the advent of large generative models, many companies no longer require the classical R&D department. Innovations come from development with additional responsibilities—taking what's already available, combining it into the desired state, and moving to production. This approach yields predictable results that are much closer to traditional software development.