MercadoLibre is the largest online retailer in Latin America with over 100M users. It is an e-commerce and auction platform that provides users with buying and selling mechanisms for e-transactions. We established a working relationship with them in order to solve issues related to the category tree of their product catalog.
E-commerce businesses spend a considerable amount of time and effort in organizing their products. MercadoLibre has a vast catalog, lots of data that needs to be archived and correctly analyzed. They wanted to explore a solution that makes that category tree more efficient and examine ways of automating this process.
An iteration based on specific goal achievement, ongoing development and progressive team scaling were the basis from which this solution was crafted. Classifying products into a set of known categories by using supervised learning.
The problem was faced in three stages: Research & Development, bringing the model into production, and exposing it through an API and Frontend development.
R&D implied a massive historical data analysis which later directed us to the hypothesis, trial & error circle until a fair solution was found.