Modulai built a recommendation system for Skincity that combines the unique knowledge of their skin therapists with product and customer behavior data. The system generates relevant brand recommendations for each skin test submission, aiming to drive higher conversion rates and customer satisfaction.
Challenge
Skincity is an online skincare clinic offering a curated selection of professional skincare products and makeup. Central to their operation are the customized product recommendations customers receive from experienced skin therapists. Many customers taking the skin test are first-time visitors, and the therapists rely on information such as age, skin type, condition, and preferences for organic or vegan products.
Solution
To assist therapists and leverage their expertise further, Modulai developed a machine-learning-based system generating relevant and accurate brand recommendations for each skin test submission. The automated recommendations, combined with the therapists' expertise, result in customized product sets and treatment plans. The team analyzed years of test submission data and purchase patterns, developing a hybrid collaborative filtering model based on user and product attributes, past interactions, and test submissions.
Tools
Models were trained using algorithms for tabular data and collaborative filtering. The backend was primarily developed in Python and deployed as a set of dockerized services. Various AWS services were used for data pipelining and scalable deployment.
Learn more
