Case study
A store with a recommendation neural network
Online store with a built-in recommendation system based on a neural network
01 Task
In an online store, customers left the product card without seeing anything similar, the “best sellers” block showed everyone the same thing and did not attract interest
02 Solution
I built a store management application in .NET and connected a recommendation neural network that selects similar products based on history and characteristics, and not on the overall top. The model was taken out separately from the interface so that it could be retrained on fresh sales without touching the main code
03 Solved tasks
- Designed and implemented the architecture of a recommendation neural network
- Connected a neural network instead of a common block of bestsellers
- Implemented a selection of similar products based on history and characteristics
- I took the model separately from the interface for retraining on fresh sales
- Combined recommendations and remainders into one manager window on .NET
04 Result
A live block “they take it with this” for a specific product appeared on the card. The manager sees recommendations and balances in one window and edits the assortment from there