Speaker Name: Dr. Soumajit Dutta

Date: 24-10-2025 (Friday)

Time: 2:30 PM

Venue: CL119

Abstract: The world is facing a multitude of urgent challenges, from the rising cost of R&D in drug development in the wake of potential future pandemics to the escalating demands of climate change, energy, and food security. Addressing these issues requires accelerated innovation in molecular and materials engineering—areas where chemical engineers play a pivotal role. Recent advances in computational algorithms and hardware have enabled both physics-based simulations and data-driven machine learning methods to emerge as powerful, cost-effective tools for high- throughput molecular discovery and design.
In this talk, he will discuss how the integration of physics-based and data-driven models is leading to more robust, interpretable, and efficient molecular-scale predictions for applications in rational drug design and quantum computing. He will highlight recent work showing how data-driven models trained on molecular dynamics simulations can accurately characterize ligand binding and reveal allosteric mechanisms in cannabinoid receptor signalling. He will also present how machine learning force fields and data-driven collective variables enable efficient modelling of defect dynamics in silicon carbide, a material of growing interest in electronics and quantum technologies. Overall, these integrated approaches hold significant promise for accelerating molecular engineering workflows across a broad range of chemical engineering applications, from pharmaceuticals to advanced materials and energy systems.