applied deep learning for natural language understanding

We find meaning in spoken, conversational input by isolating and evaluating quantitative concepts and positively identifying non-quantitative elements. Unstructured, often incomplete data is transformed into structured usable information.

our nlu systems

Our first production NLU system uses sophisticated AI models to deliver exceptionally high food identity and quantity match accuracy rates. The first of its kind specifically optimized for the food domain, our system was trained on our unique library of spoken, conversational food representations collected over more than a decade.

See it in action. Try Talk-to-Track.


With current applications in the areas of disease management, food commerce, diet and nutrition and more, we are expanding our system to offer services in healthcare and broader e-commerce.