Scientific exploration slowed by application plumbing
Solenne’s innovation studio builds digital tools for product scientists and brand teams. One application helps researchers remove or replace ingredients while preserving performance, texture, fragrance, cost and regulatory constraints.
The original application pulled from several warehouses and document repositories, then performed extensive restructuring and calculations in backend scripts. A single investigation involved many interactive steps, making even small delays visible to scientists and difficult for developers to optimise.
Formulations, evidence and AI context in one model
The team moved formulation records, ingredient properties, experiment results and regulatory references into SelectiveDB. Flexible documents represent evolving scientific structures, while vector indexes connect laboratory notes and published evidence to operational formulation data.
Selective Cloud runs the platform across research centres and automatically scales simulation workloads. Developers use database-side aggregation and retrieval instead of maintaining large amounts of bespoke calculation and synchronisation code.
More experiments in every research session
Interactive query latency fell to 12 milliseconds and formulation retrieval became 32 times faster. Researchers can now test alternatives continuously rather than submitting a request and waiting for a batch result.
The new model removed 45% of custom calculation code and indexes more than 700,000 formulation and ingredient relationships. Teams can extend the application to new brands and regulatory datasets without redesigning the database.