I’m Dario and for six months I’ve had the incredible opportunity to work at Prime Video as an applied scientist intern. When I joined, I couldn’t have imagined how much fun and impressive my personal and professional growth would be during this unforgettable experience. Looking back on my time at Prime Video, I’d like to share some of my experiences and insights for anyone who is considering internship opportunities at Amazon.
What’s it like to be an applied scientist intern at Prime Video?
Let’s start from the beginning. During my degree in computer engineering at the Polytechnic University of Turin, I became fascinated by the potential of artificial intelligence (AI) and machine learning (ML). After graduating I decided to deepen my knowledge of the field by starting a Ph.D. in computer vision. As I approached my Ph.D.’s completion, I wanted to acquire more practical experience by working on improving products while adopting ML solutions.
So, in 2022, I joined the Detection, Algorithm and Content (DAC) team at Prime Video as an applied scientist intern based in London, working with my manager Sathya Balakrishnan, software development manager, and my mentor David Higham, computer vision scientist.
The overall Video Quality Analysis (VQA) team (to which DAC belongs) owns the challenge of continuously raising the quality of Prime Video content that is released to customers. We also own the development and maintenance of automated detectors to identify audio/video detects in both the Prime Video catalog and during live events. In this context, the goal of my project was to develop an automated algorithm able to predict the quality of a never-seen-before video without any additional information.
At the start of my internship, Sathya gave me the freedom of exploring and choosing the best development environment. I researched multiple development environments and found one that combined both the flexibility I needed along with low operating costs. While planning short-term and long-term projects, Amazon scientists need to be able to design and plan how to Deliver Results, while not forgetting the Frugality Leadership Principle.
However, delivering results is only one aspect of our work. As scientists, we crave and love challenges. We are expected to tackle unsolved challenges, where neither the problem nor the solution are defined. We never take a decision for granted, and we keep on challenging ourselves to push the Amazon science bar higher. This is what we love doing and what I mostly did during my internship, where I have had the opportunity to Learn and Be Curious while exploring the state-of-the-art of video quality assessment.
While trying to break my project down into smaller steps, I initially investigated adopting several video quality assessment models for streaming defect monitoring. Because we were not bounded by any constraint in terms of speed and resource usage for this use case, I experimentally compared ChipQA with deeper state-of-the-art models to understand which direction would have brought us the highest performance.
The results of my analysis were submitted and accepted at the 4th Workshop on Video/Audio Quality in Computer Vision: Machine Learning and Beyond at the internal Amazon Computer Vision Conference 2022. Last but not least, towards the end of my project I also explored advanced self-supervised techniques, having the chance to work on cutting edge technologies.
Is it all work as an intern?
Of course not! As an intern, you’re provided with a tremendous variety and breadth of experiences to learn and grow from. Being extremely passionate about research, I attended the Computer Vision and Pattern Recognition (CVPR) conference in New Orleans, Louisiana, which is the world’s top conference for sharing and learning about cutting-edge solutions to challenging problems in computer vision and related fields.
It was an incredible experience, full of joy, fun, and learning opportunities. I met lots of amazing researchers, built connections with more experienced people and attended incredibly high-quality presentations. I also joined our internal Prime Video UK Science Conference 2022 as a speaker, where I connected with peers in London and exchanged ideas about how to further improve Prime Video products. It’s been exciting and inspiring to see how many projects, teams, and people work together to offer the best possible customer experience.
I also presented a few demos to showcase my progress to leadership and colleagues, as well as hosted a few presentations on machine learning topics in the VQA ML learning group. These opportunities helped me hone my public speaking skills and explaining science to a broad audience.
As an intern you also get to be part of an amazing community, participating in internal events and activities and growing your career while building relations with remarkable people from all over the world. One of the recent events I enjoyed the most has been taking part of the exclusive preview screening of The Lord of the Rings: The Rings of Power. It has been a memorable night out and witnessing the results of the great, passionate, and dedicated effort done by so many teams from across the world truly gave me shivers!
Interested in joining Amazon for an internship?
It’s almost impossible to sum up in a few sentences what this internship at Prime Video has meant to me. But, I will attempt to highlight the most important experiences. First and foremost, I got the opportunity to apply what I had learned during my Ph.D. to more concrete scenarios and see the impact I was able to make. I also learned a lot about how science is pursued in a corporate setting, which was very interesting when compared to how I was used to doing it in a university setting. Aside from the pure science topics, I also learned a lot about myself. I met amazing people who taught me to always push myself further, to view every challenge as a learning opportunity, and to force myself to step outside of my comfort zone. I finished my internship at Prime Video with unforgettable memories and would strongly recommend such an internship to anyone interested in science.