May 2023 – Jun 2025 · 2 yrs 1 mo · On-site

• Led the team responsible for the research and development of the Netflix recommendation algorithms using machine learning. This includes end-to-end work on personalizing rankings, rows, modules and full page layouts, as well as optimizing how content is presented across all devices. Our work spanned research, design, implementation, evaluation, deployment, and ongoing operation in these areas, leveraging both offline experimentation and large-scale online A/B testing.
• Drove cross-functional efforts to improve the algorithm, infrastructure, and product features. Championed and led cross-functional R&D initiatives to bring large language models (LLMs), generative, and adaptive techniques into the Netflix personalization experience. This included exploring new ways for models to understand member tastes and content, enable natural language interfaces and explanations, and prototype agent-based approaches.

Sep 2017 – May 2023 · 5 yrs 8 mos · On-site

• Led the team responsible for research and development of the machine learning algorithms that create a personalized Netflix homepage for our over hundreds of millions members across all devices. This spans personalizing the rows and videos recommendations as well as personalizing how we display them. We work end-to-end by researching, designing, implementing, evaluating, and deploying algorithms in these areas, through both offline experiments and online A/B testing.

Mar 2014 – Sep 2017 · 3 yrs 6 mos · On-site

• Led the Page Algorithms Engineering team of researchers and engineers. We use machine learning and other algorithmic techniques to improve personalization and recommendations on the Netflix homepage.

Jul 2011 – Mar 2014 · 2 yrs 8 mos · On-site

• Worked on research and development of personalization algorithms and recommendation systems using machine learning. Working on improving the core personalized ranking system that drives the majority of recommendations in the Netflix streaming experience, including most of the homepage.
• Researched, developed, and deployed several improved machine learning approaches that resulted in statistically significant improvements in the member experience.
• Designed and developed the core code and API for the ranking system including components for personalized feature generation, scoring, and sorting of videos. Optimized the engine to be capable of ranking entire catalog for a user in real-time and scaled it to work for millions of users. Also developed supporting tools including new machine learning algorithms and ranking quality metrics.
• Designed and developed core machine learning algorithms and supporting tools used across multiple production personalization and recommendation algorithms.

Sep 2004 – Jun 2011 · 6 yrs 9 mos · Albuquerque, NM · On-site

• Led the design and development of the Cognitive Foundry, a software library for machine learning and cognitive simulation.
• Led a team researching incremental (online) machine learning algorithms.
• Conducted research in large-scale machine learning in map-reduce (Hadoop) using ensemble methods.
• Led the research, development, and deployment of personalization technology for enterprise search.
• Led the design, development, and deployment of an email categorization system that combined text analysis and machine learning, which was successfully deployed.
• Co-developed an augmented cognition system for detecting difficult driving situations due to both driver overload and underload, in conjunction with a commercial partner.
• Advocated and led the creation of the department’s software engineering processes.

Jan 2002 – Jan 2004 · 2 yrs · Providence, RI · On-site

• Master’s research project in machine learning and information retrieval: a machine learning approach to recommender systems by unifying collaborative filtering and content-based filtering.
• Used a kernel-based machine learning approach to couple together learning problems and integrate all available information.
• Extensive experiments show improvement over state-of-the-art methods on a movie recommendation task.

May 2001 – Aug 2001 · 3 mos · El Segundo, CA · On-site

• Helped design and implement a system of distributed, intelligent agents in a simulated environment that use emergent behavior to detect and use resources to automatically accomplish joint goals and fix failures in the underlying communications network.