Meals App

Challenge

The meals.app team contacted us to help them with improving their recipe exploration features by leveraging machine learning techniques like recommendation systems and natural language processing.

Solution

First we expanded the recipe database by collecting and synthesizing new data. Then the recommendation system was extended in to more features and improved recommendation quality. Lastly, we improved search functionality by using embeddings and other NLP techniques.

Results

• Added ~15000 new recipes to the meals.app database
• Improved user engagement with suggested recipes by +30% with the potential to improve healthy eating habits
• Expanded search and recommendation capabilities with new tags like ‘healthy’ and ‘calorie dense’ better helping users with their diet goals
• Improved the relevance of search results

Meals App