Machine Learning Through Neural Networks – My InStep Project Trajectory

As an InStep intern, I worked on the project to automate product tagging and categorization using machine learning technology. The project team guided me to build a web-scraping tool to collect data from multiple popular ecommerce sites and to analyze and clean the data using web scraper.

Gabriel Brusman
University of California, Davis, USA

Gabriel Brusman

In my project, I implemented a multi-class, multi-output neural network that could learn many classes of different attributes about specific product types, and then assign images of products to a class for each attribute. Since there were many different attributes and many classes for each attribute, the end result was a detailed categorization of each image. For example, it could tell you that a shirt was a red, casual, short-sleeved t-shirt with a V-neck. I also made a small dashboard to show off my project and the value that it brings to the company.

During the 12 weeks at Infosys, I was able to contribute to an end-to-end machine learning project, while interacting with professionals in engineering and data science streams. I gave a few different presentations and tech demos about my project to people from the analytics department all the way to the marketing department, along with exploring India and immerse myself in its culture and form new connections with fellow interns from all over the world. It was truly a summer I will never forget.