Inside Hulu’s AI Research: Personalization, Data Science & Video Innovation

Workshop Announcement

Hulu’s second PhD workshop will be held on December 29 (Thursday) from 6:00 pm to 8:00 pm. Interested students should send their résumé (or LinkedIn file) to [email protected] and mark the subject with “PhD Workshop”. Invitation emails will be sent by the end of December 28.

About Hulu

Founded in 2007, Hulu is a U.S. video‑streaming platform backed by Disney, 21st Century Fox, Comcast and Time Warner. It aims to provide the best viewing experience for high‑quality video content, serving over 11 million registered users and generating more than $2 billion in annual revenue.

The company places great emphasis on its technology teams, operating research centers in Los Angeles, Seattle, New York, Chicago and Beijing. The Beijing team, in particular, conducts algorithmic research based on large‑scale content and user data, influencing Hulu’s strategic decisions.

Hulu’s research culture promotes innovation, collaboration, and external engagement, encouraging rapid adoption of leading‑edge findings and fostering a supportive, trust‑based environment among researchers.

Research Areas

Personalization & Content Discovery : The team strives to deliver a tailored viewing experience by deeply understanding users. Ongoing projects include scenario‑aware recommendation, personalized algorithms, and the “A neural autoregressive approach to collaborative filtering” (ICML’16) which achieved state‑of‑the‑art performance on public datasets.

Data Science : Leveraging Hulu’s massive content, user, and operational data, the team extracts actionable insights for strategy, product, and market decisions. Current projects cover user churn analysis, content clustering, and large‑scale behavior understanding.

Video Codec & Optimization : Research focuses on next‑generation video codecs, adaptive bitrate algorithms, QoS/QoE optimization, and efficient video distribution and network optimization to ensure high‑quality streaming across devices and environments.

Content Understanding : By extracting rich metadata (people, emotions, events, dialogue, music) from video programs, the team enables automated video description, supporting recommendation, search, and data‑science pipelines, and explores interactive innovations for more natural user experiences.

Intelligent Search : The team tackles the challenge of fast, accurate search across devices, aiming to understand user intent and support multiple retrieval modalities such as text and voice.

Advertising Intelligence : Using machine intelligence, the team improves ad services through precise targeting, traffic forecasting, pricing, and flow management, enhancing Hulu’s significant ad‑revenue stream.

Workshop Topic 1 – Statistics Overview

“Statistics is a branch of mathematics concerning data collection, analysis, interpretation, and presentation. This workshop covers fundamental concepts of frequentist statistical inference, including point estimation and hypothesis testing.”