As pharmaceutical entities, and the processes used to produce them, continue to grow in complexity, data-rich technologies are increasingly playing a critical role in building process understanding, especially under short development timelines. This talk will showcase recent advancements by the Data-Rich Experimentation (DRE) group at Merck, a multidisciplinary team advancing process development through high-throughput experimentation, custom automation, in silico development and digital innovation. We will present a few case studies that exemplify our approach, including the broad deployment of algorithmic process optimization enabling real-time reaction tuning, reduced cycle times and improved resource utilization. Second, we’ll discuss our hybrid modeling framework, GARNET, which combines mechanistic models with neural ODEs to simulate complex reaction dynamics, accelerating scale-up and enhancing predictive control, particularly in the upstream bioreactor space. Third, we’ll demonstrate how large language models have empowered scientists to rapidly prototype custom instrumentation and user interfaces, greatly streamlining automation activities. Together, these innovations illustrate how the DRE group is redefining process development— delivering faster, smarter, and more resilient experimental technologies.
