Senior Research Scientist - Discovery (Computer Vision)
- Los Gatos, California
- Science and Analytics
Netflix is seeking a passionate and talented machine learning expert with a strong background in computer vision to help portray the shows in our catalog and make them resonate with our subscribers. In a role that sits at the intersection of computer vision and artistic creativity, you will investigate, conceptualize, design, implement, and validate new algorithms that will help curate our visual assets such as artwork and trailers and harness them within the product for a personalized experience. Rather than just use computer vision to improve accuracy on a task, leverage it to drive real engagement across hundreds of millions of users!
To learn more about our research and analytics work, you can visit our research page here.
As a senior research scientist, you will:
- Work closely with various artwork design teams and product development engineering teams to solve key asset optimization, personalization and discovery problems.
- Translate unstructured, complex business problems into an abstract mathematical framework, making intelligent approximations when needed to put your algorithm to work at scale.
- Bring a combination of mathematical rigor and innovative algorithm design to create recipes that extract relevant insights from billions of rows of data to meaningfully improve user experience.
- Join a world-class team of machine learning researchers with an extensive track record in both academia and industry.
- Interact with and report to an audience that includes Directors, Vice-Presidents and the Chief Product Officer.
Some examples of the problems you might tackle in your new role:
- How can we personalize the use of visual assets for our content based on user engagement? Artwork Personalization at Netflix
- What is the recipe for creating engaging visual assets that can help make our content relevant and compelling to our audience? Early insights into the problem: The power of a picture
- How do we mine our rich content library and categorize the scenes in support of an immersive and interactive user experience? Learnings from related efforts: Extracting image metadata at scale
- How can we determine which specific asset from a candidate pool for a show will drive the highest engagement for that show? An example of some of our efforts on this question: Selecting best artwork for videos through A/B testing.
Netflix is a fact-based, analysis-driven organization:
- The improved algorithms you develop will be A/B tested quickly and rigorously and you will have direct, measurable impact to the user experience and the bottom-line.
- MS or PhD degree in Computer Science, Statistics or related field (PhD preferred)
- Strong background in computer vision using unsupervised and supervised methods for semantic analysis and scene understanding of videos and images with a track record of publishing in top tier journals and conferences such as NIPS, CVPR and ICCV.
- Proven track record of leveraging large amounts of data to solve real-world problems.
- Experience with recommender systems, contextual bandits, online learning, reinforcement learning, causal learning would be a plus.