Title
Sustainable Digital Chemistry: Scaling ML-Driven Workflows for Energy-Efficient Discovery
Abstract
The urgency of reducing computing's environmental footprint makes it critical to embed sustainability at the heart of scientific software and infrastructure. In this lecture, Dr. Jan Gerit Brandenburg will present strategies for constructing energy-aware, ML-assisted computational chemistry workflows that can transition from academic prototypes to efficient, industrial-scale platforms. He will discuss how to optimize algorithm choice, dataflow, model architectures, and infrastructure (cloud, HPC, hybrid) in ways that balance scientific throughput with power and carbon costs. Through case studies - ranging from impurity detection in OLED manufacturing to closed-loop autonomous synthesis labs - he will show how digital chemistry can evolve into a greener, scalable engine for innovation in materials, life sciences, and process R&D.
Speaker
Dr. Jan Gerit Brandenburg is Director of Digital Chemistry at Merck, where he leads the development of scalable, sustainable computational solutions for molecular and materials innovation. With over a decade of experience bridging academia, industry, and entrepreneurship (co-founder & CEO of Quastify), his work spans quantum chemistry, molecular simulation, and ML-driven models applied to real-world R&D challenges.
He holds a PhD in theoretical chemistry and has published broadly on the intersection of computational science, machine learning, and scalable workflow architectures.
Using Data Science Methods Such as Active Learning and Gaussian Process Models to Minimize Simulations and Speed Discovery
Edward Maginn (pronunciation)
Keough-Hesburgh Professor &
Associate VP for Research
Notre Dame, IN 46556 USA