Greater New York City Area
Google Labs: Action Intelligence
Mission: Teach large language models (LLM) new tricks
Core skills:
1. Generative AI: advanced RAG, prompt/model/embedding tuning, reranking, grounding, multi-step reasoning, agentic tool-use
2. Planning AI: deep Q-learning, Monte Carlo tree search
3. Frontend: Angular, Lit, Material Design
4. Midtier: microservice architecture
5. Backend: continuous and batch processing
6. Cross-team coordination for horizontal changes that touch across many Google products
Patent:
1. Efficient Accumulation of Agent Behavior Statistics, GP-301709-00-PCT (IDF-300053). Basically, we generalized hyperloglog from doing count(distinct categorical_values) to sum(distinct numerical_vectors) in order to support an application of machine learning on agent behaviors.
Publication:
1. https://research.google/pubs/pub50207/
Key projects:
1. Research prototype web-agent Project Mariner https://deepmind.google/project-mariner
2. Author of Semantic Retrieval API in LlamaIndex and LangChain.
3. Brand-Safety frontend for Campaign Manager 360 and Display & Video 360.
4. Access-analysis to measure Google-wide data-protection posture of RPC endpoints.
5. Reach & Frequency, Gross & Target Rating Point (GRP & TRP). Brand-advertising strategy. Inventory Management & Forecasting.
6. Google's next generation stateful streaming data processing infrastructure.
Other positions taken at Google:
1. Database and Storage System. How large volume of data is stored and maintained?
2. Inventory Management and Forecasting. How to predict how many impressions would be served?
3. Ad Manager Reporting. How to build really large scale data processing pipelines?
4. Unified Reporting Infrastructure. How to build really low latency serving systems?
5. Data Protection. How to secure data access and measure user privacy?
6. Ads Verification Frontend. How to build beautiful, user-centric workflows?
