Kumar Rishabh
Tech Lead / Senior ML Engineer at Google — Workspace Intelligence
San Francisco, CA
I'm a Tech Lead / Senior ML Engineer at Google, currently working within Workspace Intelligence, which builds agentic intelligence and personalization for Gemini. I focus on evals and harnesses, agentic systems, and the self-improvement loops that feed into Gemini's post-training.
I've spent almost a decade at Google, previously leading ML for YouTube Ads and founding the Sports product in Google Search. My background is in distributed systems and cloud storage research (M.S. by Research, IIIT Hyderabad), which still shapes how I approach large-scale ML infrastructure today.
Selected Work
Tech Lead, YouTube Ads ML
Led ML for YouTube Ads, an incremental multi-billion dollar ARR business, across a team of about ten engineers.
- Led LLM-based generative retrieval, recommendation, and ranking models.
- Owned privacy-preserving ML, leading model and systems development including transfer learning for ad-targeting.
- Built bidding models and brand / direct-response audience models for personalized, contextual targeting.
- Recipient of a Google Tech Impact Award for this work.
Founding Engineer, Google Sports Search
Founding engineer of Sports on Google Search, Feed, and Notifications.
- Led efforts to improve sports search quality, discoverability, and ranking.
- Conceptualized and built 50+ features spanning search experience, data quality, and freshness.
- Built multi-modal experiences like live cricket and US sports notifications, and "where to watch" recommendations.
Google Research
Improved search result quality and recommendation relevance in Google Play.
Software Engineering Intern, Meta
Interned on the Network Research team, building software-defined networking (SDN) systems for Meta's datacenter network.
Patents in Machine Learning, Generative AI & ML Infra
Multiple patents issued spanning machine learning, generative AI, and ML infrastructure for content serving.
- Privacy-preserving machine learning and differential privacy techniques.
- Generative dataset and configuration pipelines for content serving systems.
- Prediction calibration and resource-efficient ML infrastructure for content delivery.
Machine Learning for Cloud Storage Systems
My M.S. by Research thesis applied machine learning and statistical analysis to heterogeneous cloud storage systems.
- Multiple peer-reviewed publications in distributed systems and cloud storage.
- I was a repeat Google Summer of Code student and Linux Foundation intern, mostly working on SDN and distributed systems projects.