Cristos Goodrow
Cristos Goodrow is an engineering leader whose AI work spans YouTube search and recommendation systems. He studied mathematics and computer science at Cornell after first hoping to become an architect, then earned a mathematics master's degree at UC Santa Barbara and pursued doctoral work at UC Berkeley. At YouTube, he led Search and Discovery, explained the system's use of watch time and similar viewing habits, and coauthored work on recommendation fairness, fresh content, and language models. In February 2026, he testified as YouTube's vice president of engineering.
Timeline
-
1991
Cornell mathematics and computer science
In 1991, he graduates from Cornell after majoring in mathematics and computer science, having first hoped to become an architect. He earns a master's degree in mathematics at UC Santa Barbara and pursues doctoral work at UC Berkeley.
-
2012–2017
Watch time at YouTube
In 2012, YouTube adds watchtime to its recommendation signals to track how long viewers actually watch after clicking. In February 2017, Goodrow reports that viewers around the world watch one billion hours of YouTube each day.
-
2018–2020
Search and Discovery at YouTube
In 2018, he is YouTube's vice president of engineering and leads Search and Discovery; before YouTube, he directs Google Shopping and Amazon search-engine software. In January 2020, he is a Google vice president of engineering and head of Search and Discovery at YouTube.
-
2019
Fairness in recommendation rankings
In 2019, he coauthors the KDD paper “Fairness in Recommendation Ranking through Pairwise Comparisons,” which evaluates fairness in recommender rankings using pairwise comparisons.
-
2021
Explaining YouTube recommendations
In 2021, he explains how YouTube compares a viewer's habits with those of similar viewers to suggest other videos, including topics the viewer has not watched.
-
2023
Fresh content recommendation
In 2023, he coauthors “Fresh Content Needs More Attention: Multi-funnel Fresh Content Recommendation,” a KDD paper on a dedicated stack for fresh content. Its multi-funnel system combines a two-tower model for coverage with a sequence model that updates from user feedback in near real time.
-
2025
Language models for recommendations
In 2025, he coauthors “PLUM: Adapting Pre-trained Language Models for Industrial-scale Generative Recommendations,” a framework for adapting pre-trained language models to industrial recommendation tasks. The paper describes training and inference methods for launching it to billions of YouTube users.
-
February 2026
YouTube engineering vice president
In February 2026, he testifies in a Los Angeles social media trial as YouTube's vice president of engineering.
Career
Education
- 1991Cornell University
- University of California, Santa BarbaraMaster’s degree in mathematics
- University of California, BerkeleyDoctoral study
Sources and credits
Photo: We have no published photo of Cristos Goodrow yet. Send us a free-licensed photo.
- Alumnus guides YouTube's search
- Career Conversation - Cristos Goodrow '91, Vice President of Engineering at YouTube
- On YouTube's recommendation system
- You know what's cool? A billion hours
- Cristos Goodrow: YouTube Algorithm
- Fairness in Recommendation Ranking through Pairwise Comparisons
- Fresh Content Needs More Attention: Multi-funnel Fresh Content Recommendation
- PLUM: Adapting Pre-trained Language Models for Industrial-scale Generative Recommendations
- Youtube Executive Testifies in Social Media Trial
- YouTube exec says goal was viewer value not addiction - eNCA



