Not Your Average Internship: How My Customer Pitch Lead to Adobe Primetime’s Video Recommendations Engine

Not Your Average Internship: How My Customer Pitch Lead to Adobe Primetime’s Video Recommendations Engine
Adobe Products Featured

photo_GangWu-fimg

Companies that offer streaming video services struggle to provide subscribers with relevant, personalized recommendations. With so much content available, it can be frustrating to find programming that matches people’s interests.

During my multiple internships at Adobe, my challenge was to address that issue by helping create technology that generates personalized video recommendations. We replaced the traditional model, based on manual user feedback, with an automated system. Today, the algorithm I worked on, powered by Adobe Target, allows customers of our multiscreen TV platform (Adobe Primetime) to use machine learning to intelligently surface data-driven, personalized video recommendations that play back instantly and highlight the videos most likely to appeal to consumers.

Of course, it didn’t happen in one summer. This was the culmination of three years interning under the guidance of Viswanathan (Vishy) Swaminathan in Adobe Research.

It all started when my research idea and the results were pitched to an Adobe Marketing Cloud customer to automate video recommendations based on tracking users’ historical viewing activities. Based on the positive response from customers, the Adobe product team decided to implement my idea. It was exciting on so many levels: not only was I one of the first interns whose idea was  pitched to a customer, it actually became a new solution capability!

The solution we developed combines a user’s engagement and content consumption history with contextual information – such as the consumer’s device, profile, geographic information, and aggregated viewing data that reflects overall viewing preferences.

This technology, enabled by Adobe Target, is the backbone of Adobe’s new video personalization engine – Adobe Primetime Recommendations. It leverages automated data insights around individual and aggregated video consumption to increase viewer engagement and viewing time. Announced at Adobe Summit 2016, Adobe Primetime Recommendations learns from more than 200 billion online video consumption points across most U.S. households that stream TV and film content.

It’s really cool that I could contribute to a real-world solution while working as an intern. Adobe is a company where your ideas can ultimately have an impact on the industry. Plus, it’s a strong company — and it’s located in the center of the technology universe.

My internship provided me with invaluable product development experience, from understanding a problem to identifying tools and methodologies for developing solutions.

Plus, working with experienced data scientists has helped fine-tune my analytical skills, and enabled me to apply both my engineering and research skills. Vishy’s instruction and guidance has been very important in all the work I have done here.  Even though I’m still finishing up my Ph.D, I have already worked on four research papers and have submitted applications for two patents.

Much of this is fueled by having access to cloud computing and big data from Adobe Marketing Cloud– which supports more than 41 trillion transactions per year on behalf of its customers worldwide! It’s an experience you just can’t have on a college campus.

Looking ahead, I expect to finish up at Iowa State University this year and, hopefully, Adobe Research will be part of my future.

Some of our most amazing tech advancements likely started with an intern or university collaboration project. Check out some of the other stories in this series, which share how interns are the secret to a thriving research lab.

You may also like:

Recommended Articles