Discover is Google’s personalized interest feed, connecting users globally with the best content from across the web in real time.
Every Discover feed starts with Discover User Modeling and Retrieval. Because core retrieval and ranking rely heavily on historical engagement, newly ingested documents remain corpus "dark matter" and expire before reaching the signal thresholds needed to show their quality. We built Discover’s end-to-end content exploration system to break this bottleneck, spanning real-time multi-tier lifecycle control and embedding-based cold-start retrieval.
global scale. We apply breakthrough research, leveraging models like Gemini, to launch high-impact products that help billions ask any question and get things done. If you want to drive the future of AI in Search and work with world-class experts to deliver tangible, global impact and continuous innovation, this is the place.Individual pay is determined by factors including job-related skills, experience, and relevant education or training.
US: $174000 - $252000 (USD) + 15% bonus target + equity + benefits
Learn more about benefits at Google.
Responsibilities
- Research and deploy ML models that identify content by combining multimodal LLM understanding, uncertainty estimation, and causal modeling to maximize exploration ROI at scale.
- Drive projects from problem formulation and experimentation design through model training, real-time serving architecture, and production launches, collaborating closely with Feed Ranking, Content, and Infra teams.
- Partner to align semantic applicant retrieval, ranking incentives, and warm-start handoffs so promising new content seamlessly activates the organic recommendation flywheel.
- Lead zero-history discovery for Discover’s top formats—making dynamically generated AI content (AIGC) a citizen across the stack and surfacing breaking news within minutes of publication.
- Evolve StarSearch into the intelligent gatekeeper for Discover’s content ecosystem—safely bootstrapping fresh, niche, and creator supply to expand personalization headroom while rapidly pruning unengaged content to keep the serving corpus fresh and efficient.
Minimum qualifications:
- Bachelor’s degree or equivalent practical experience.
- 5 years of experience with software development in one or more programming languages.
- 3 years of experience with end-to-end Machine Learning (e.g., model deployment, model evaluation, data processing, debugging).
- 3 years of experience building and deploying recommendation systems models (retrieval, prediction, ranking, personalization, search quality, embedding) in production; and experience building architecture.
- Experience developing personalization algorithms for user-specific content delivery.
- Experience integrating generative AI tools or LLM interfaces into workflows.
Preferred qualifications:
- Master's degree or PhD in Computer Science, or related technical field.
- 5 years of experience with data structures and algorithms.
- Experience in Content Exploration/Exploitation (e.g., Bandit systems, Reinforcement Learning) or solving Cold Start problems.
- Experience with feed recommendation systems or budgeting systems.
- Experience with data pipelines and system design for real-time evaluation.