ZeroShotAPI: Using ZeroShot Machine Learning and an automated HTE reactor to derive universal chemical reaction condition parameters for Active Pharmaceutical Ingredients.

Making a new drug is like trying to bake a perfect cake when you don’t know the oven temperature, baking time, or even which ingredients to combine. This project builds a software platform that uses machine learning and an automated chemical reactor to figure out the ideal conditions—temperature, solvent, catalyst—for manufacturing active pharmaceutical ingredients (APIs) in a single shot, without trial-and-error guesswork. Currently, chemists often run hundreds of small experiments to find the right reaction conditions for each new drug molecule. That process is slow, expensive, and generates chemical waste. The platform aims to replace this with a universal set of parameters that work across many different reactions, learned from high-throughput experiments and proprietary data. If successful, the platform could slash the time and cost of bringing a new drug to market. It would also make pharmaceutical manufacturing more sustainable by reducing the solvents, energy, and raw materials wasted in failed reactions. For the public, this means cheaper medicines and a smaller environmental footprint from the factories that produce them. The project is applied, not fundamental science—it directly targets a bottleneck in industrial drug production.