Skip to content
CodeBranch
Marketing

Recommendation System based on AI

The CodeBranch's dedicated development team built an AI recommendation system for a B2B Marketplace.

Quick Summary

  • ▸ The CodeBranch's dedicated development team built an AI recommendation system for a B2B Marketplace.
  • The recommendation system is used to recommend products to customers that best meet their needs
  • The recommendation system operates in 32 dimensions for high-precision personalization
Tech Stack: Python Scikit Torch
Recommendation System based on AI

Overview

CodeBranch's dedicated development team built an AI recommendation system for a B2B marketplace. The recommendation system is used to recommend products to customers based on the specific needs of the customer and the unique characteristics of that customer. To make the recommendation, the system compares this information with that of other customers who have purchased similar products, who have similar characteristics to the current customer, and who have also provided positive reviews and recommendations of the requested products.

Industries

Marketing Hospitality E-commerce

Services Provided

  • Web Development
  • Custom Software Development
  • AI Development

Approach

The team added a Python microservice running Scikit and Torch. Data is ingested directly from the database to create a plot of points on a 32-dimensional universe. A distance-based algorithm (KNN) combined with an "ideal point" strategy finds the closest neighbors to each point. Candidates are then ranked with a normalized distance score and fed to a neural network to decide the top 5 products that may be the best fit for the requesting user (represented by the ideal point). The amount of preprocessing for a 32-dimensional universe was intense, since everything needed to be normalized and standardized. The project involved 3x senior, 1x mid-senior, and 1x junior engineers.

3x Senior Engineer
1x Mid-Senior Engineer
1x Junior Engineer

Results

  • The recommendation system is used to recommend products to customers that best meet their needs
  • The recommendation system operates in 32 dimensions for high-precision personalization

Find Out More

Project Home Page

Frequently Asked Questions

How does the recommendation engine personalize results for each customer?
The system builds a profile for each customer based on their specific needs and unique characteristics, then compares that profile against other customers who have purchased similar products, share similar traits, and have left positive reviews. This multi-factor comparison produces a ranked list of the most relevant products.
Why was a 32-dimensional model chosen for this recommendation system?
A 32-dimensional space allows the system to encode a rich set of customer and product attributes simultaneously, enabling far more precise similarity matching than lower-dimensional models. This depth of representation is what allows the system to surface products that genuinely fit each customer's unique context.
What algorithms power the recommendation engine?
The system uses a K-Nearest Neighbors (KNN) distance-based algorithm with an "ideal point" strategy to identify candidate products, then ranks those candidates using a normalized distance score. A neural network (built with Torch) makes the final selection of the top 5 recommended products.
How did the team handle the large volume of data involved?
Data preprocessing was one of the most intensive aspects of the project. The team built a robust pipeline to normalize and standardize all inputs before ingestion into the 32-dimensional model, ensuring consistent and reliable recommendations despite the scale of the dataset.
Can this type of AI recommendation system be applied to B2C marketplaces as well?
Yes. While this system was built for a B2B context, the underlying architecture — combining KNN similarity search with neural network ranking in a high-dimensional space — is equally applicable to B2C e-commerce, content platforms, and any use case requiring personalized recommendations at scale.

Related Case Studies

Review and Reputation Marketplace Company

Marketing

Review and Reputation Marketplace Company

E-commerce Solution with Smart Assistant

E-Commerce

E-commerce Solution with Smart Assistant

AI Agent to Optimize Decision-Making in Supply Chain Planning

Supply Chain

AI Agent to Optimize Decision-Making in Supply Chain Planning

Ready to Build Something Great?

Let's discuss your project and find the perfect solution for your business.

Start a Project