Amey Dharwadker, a Machine Learning Engineering Manager at Facebook and leader of the Video Recommendations Quality Ranking team, discusses the complexities of personalizing video feeds for millions of users. He highlights the challenges of real-time personalization in fast-paced content environments and the cold start problem with billions of videos. Amey also delves into the significance of user engagement metrics and cross-domain data in refining recommendations, aiming to create diverse and meaningful viewing experiences.
01:21:20
forum Ask episode
web_stories AI Snips
view_agenda Chapters
auto_awesome Transcript
info_circle Episode notes
insights INSIGHT
Videos Dominate Facebook Usage
Videos constitute over 60% of global time spent on Facebook, highlighting their major importance on the platform.
Facebook has multiple video recommendation surfaces, including newsfeed, video channels, and a unified video tab.
insights INSIGHT
Balancing Long and Short Videos
Ranking both long and short videos together requires debiasing engagement metrics like watch time.
Length-bucketized engagement is used to compare a user's interest relative to similar video lengths.
volunteer_activism ADVICE
Leverage Content for Cold Start
Use content understanding extensively for cold start videos without engagement data.
After videos gain engagement, user interaction data better represents content relevance.
Get the Snipd Podcast app to discover more snips from this episode
In episode 24 of Recsperts, I sit down with Amey Dharwadker, Machine Learning Engineering Manager at Facebook, to dive into the complexities of large-scale video recommendations. Amey, who leads the Video Recommendations Quality Ranking team at Facebook, sheds light on the intricate challenges of delivering personalized video feeds at scale. Our conversation covers content understanding, user interaction data, real-time signals, exploration, and evaluation techniques.
We kick off the episode by reflecting on the inaugural VideoRecSys workshop at RecSys 2023, setting the stage for a deeper discussion on Facebook’s approach to video recommendations. Amey walks us through the critical challenges they face, such as gathering reliable user feedback signals to avoid pitfalls like watchbait. With a vast and ever-growing corpus of billions of videos—millions of which are added each month—the cold start problem looms large. We explore how content understanding, user feedback aggregation, and exploration techniques help address this issue. Amey explains how engagement metrics like watch time, comments, and reactions are used to rank content, ensuring users receive meaningful and diverse video feeds.
A key highlight of the conversation is the importance of real-time personalization in fast-paced environments, such as short-form video platforms, where user preferences change quickly. Amey also emphasizes the value of cross-domain data in enriching user profiles and improving recommendations.
Towards the end, Amey shares his insights on leadership in machine learning teams, pointing out the characteristics of a great ML team.
Enjoy this enriching episode of RECSPERTS - Recommender Systems Experts. Don't forget to follow the podcast and please leave a review
(00:00) - Introduction
(02:32) - About Amey Dharwadker
(08:39) - Video Recommendation Use Cases on Facebook
(16:18) - Recommendation Teams and Collaboration
(25:04) - Challenges of Video Recommendations
(31:07) - Video Content Understanding and Metadata
(33:18) - Multi-Stage RecSys and Models
(42:42) - Goals and Objectives
(49:04) - User Behavior Signals
(59:38) - Evaluation
(01:06:33) - Cross-Domain User Representation
(01:08:49) - Leadership and What Makes a Great Recommendation Team
Recsperts - Recommender Systems Experts