For a world-leading producer of after-glow material, we devised a custom Bayesian Optimisation algorithm for their specific manfucaturing process.

Over the initial 5 batches of 40 experiments, our model steadily improved its understanding of the chemical space, yielding previously unknown recipes solutions.

Our model helped identify a solution space that was deemed theoretically impossibly to access, marked on the right.

Challenging the status quo and reaching beyond the current knowledge supported the scientific decision making and allocation of research budget, leading to higher performance and better product for the client.