Adobe

Building for retail counterfeit item detection

The summer after my junior year of high school, I was fortunate enough to do an internship at Adobe.

The wonderful Adobe office in Noida, India

I was given the opportunity to work as an Engineering Intern in the Media and Data Science Lab at Adobe. The predominant focus of the work here was to extend the applications of AI beyond Adobe’s creative suite of applications (i.e. the apps we all know and love such as Illustrator, Photoshop, and Premiere Pro). The lab worked to publish cutting-edge research as often as possible and there was a lot of cross-collaboration with Adobe’s other research teams from around the world.

The project I worked on involved building a prototype for a system capable of counterfeit item detection in retail items. For industry partners of Adobe software such as Macy’s, this is extremely important, especially when considering that many of their items end up getting sold in a third-party setting or even at unauthorized retailers overseas. I built a web scraper capable of building retail item datasets, and then built an iOS application in Swift that housed an ML model trained with Keras that had basic counterfeit item detection capabilities. I also got the chance to do market research and competitive analysis to validate the necessity for such a project, which was nothing but an idea prior to starting my internship.

One of the highlights of this experience was gaining mentorship from the Principal Scientist (now Senior Principal Scientist) of the Media and Data Science Lab, Balaji Krishnamurthy. Balaji took the time to provide guidance and expertise on the projects I was working on, as well as sharing general career advice for someone that was - especially at that time - fairly unaware of what the professional world of ML looks like, especially with an industry-specific lens. This was a very formative experience for me. It shaped a lot of my academic choices and career decisions and gave me valuable lessons in building relationships with people that are far older and more experienced than me.