Pando Bioscience is using generative AI paired with an ultra-high-throughput screening platform to accelerate enzyme engineering for industrial and pharmaceutical applications. CEO Will Cao said the system enables screening of enzyme variants at roughly 1,000-fold scale beyond traditional approaches, with results returned in days rather than months. The platform assigns performance scores to related enzyme variants and also surfaces “bad behavior” patterns, which the team then validates in wet-lab experiments and feeds back into its AI models. Cao described the company’s strategy as customer-driven, starting from a practical bottleneck—customers often know a reaction but lack a suitable enzyme or need improved activity. Pando’s portfolio spans 23 enzymes across 11 classes, indicating the platform is designed for reuse across multiple substrate types, including small molecules and nucleic acids. For drug discovery and biomanufacturing stakeholders, the core impact is speed-to-variant discovery and iterative learning loops for enzyme function optimization.