Infinity pathfinder is an algorithm-based product that uses AI to design internal trainings for Roche. Ensuring that employees keep upgrading their knowledge has always been a priority at Roche, but thanks to Infinity Pathfinder, devising internal training paths and sessions now takes seconds rather than days.

Designing efficiency

Infinity Pathfinder was designed as a self-service model that functions 24/7 without human intervention, to serve an affiliate network of employees in all global time zones. Adventure Club designed the solution together with Roche, starting from creating a clear product vision all the way to delivering an MVP. 

The Infinity pathfinder algorithm is capable of understanding complex scenarios, can deliver over 1,700 different results and saves an average of 56 work hours per designed training.

The configurations required for the AI engine to function include a large number of complex learning requirements such as past and future role, degree of expertise in the field of training, product version, training experience etc. As an adaptable mechanism it is scalable and evolves as Roche integrate new tools and software for its team and clients.

Technical approach

In order to quickly validate filtering/searching/ranking of learning modules through metadata we used ready-made search engine Algolia. This helped us identify missing metadata and lets us experiment with the results in a limited user interface. The data was then used to clarify and fine-tune the MVP scope and plan. After the validation phase we built the MVP as a website using React. This website communicated with a lightweight programming interface (API) that connects to the data module, Algolia, and other necessary services.