We are delighted to welcome you to our AIchemy Hub’s monthly webinar series.
This month’s talks:
Talk Title: Bridging Molecular Design and Synthesis: AI and Automation for Reaction Optimisation
Achieving efficient and sustainable chemical synthesis requires bridging molecular-level understanding with data-driven process optimisation. This seminar presents an integrated framework that connects molecular representation, reaction modelling, and autonomous experimentation, illustrated through applications in pharmaceutical manufacturing.
We first enhance reaction yield prediction by embedding physically meaningful molecular descriptors and reaction context, enabling models to capture both intra- and intermolecular interactions. We then extend this framework to model solvation effects on reaction kinetics, where explicitly incorporating transition-state structures provides mechanistic insight into solvent-mediated reactivity. In parallel, we model single-component solubility with explicit temperature dependence, by establishing a foundation for future mixture and formulation predictions relevant to pharmaceutical processes. Finally, we highlight ongoing work on automated amide synthesis using robotic platforms, aiming to close the loop between computation, experimentation, and sustainable process development.
Talk Title: Screening first, AI on top.
High-throughput virtual screening is a method gaining quick traction in the world of materials discovery, particularly for organic electronics. It finds its origins in drug discovery where simple cheminformatics can reign supreme; however, to build vast datasets of optoelectronic properties required the advent of advanced hardware and (quantum chemical) code. The method is a simple funnel — look at exhaustive sets and filter down to the most promising candidates based on more and more stringent criteria. This is often done with de novo material libraries for given scaffolds for a given property. We go further by asking: what if we screened the whole synthetically-available chemical space? This way, you can screen for multiple properties across the breadth of chemistry that exists today and determine the limit of exciting discoveries. This talk will go into the discoveries made through screening both with theory only and combined theoretical-experimental approaches. We will also consider: once we have the data, how can ML & AI be used for i) improving the discovery engine, and ii) bringing digestible insights to researchers who want to instantly discover new materials?
Following the presentations, there will be time for questions from the audienceWe are delighted to welcome you to the first talks in the AIchemy Hub’s monthly webinar series.
Speakers
Dr. Lauren Ye Seol Lee
Assistant Professor – UCL
Omer Omar
CEO, Apoello
