Long Yang

I’m an AI Architect and Senior Software Engineer with 14+ years building and scaling production ML systems for enterprise and consumer products, from Google and Uber to Foursquare. I specialize in LLM/GenAI and agentic platforms—especially multimodal systems, RAG/grounded retrieval, and tool/function-calling workflows—deployed on cloud-native infrastructure.

At Google, I built production agent services using grounded RAG (Vertex AI Search datastores) to deliver reliable, citation-able answers for operational decision support. I architected agent orchestration with tool/function calling, structured outputs, and workflow state management using LangChain plus internal frameworks, delivering production-grade NestJS APIs with retries, clear interfaces, and observability.

I lead high-ambiguity initiatives end-to-end and consistently drive measurable outcomes. For example, I architected an enterprise multimodal agent platform that automated manufacturing QC and operator workflows, improving accuracy by 87% and reducing manual inspection time by 65%, while establishing a GenAI quality loop (golden datasets, regression suites, retrieval precision/recall checks, hallucination triage) and governance guardrails for regulated workflows.

Previously, I engineered ranking, discovery, and personalization at Foursquare, improving engagement 23% and relevance/accuracy while shipping scalable pipelines (Kafka → Spark → MongoDB). At Uber, I delivered rider churn prediction and real-time payment fraud detection services, reducing churn 10% and fraudulent transactions 27% through robust ETL, feature pipelines, and monitoring; I also design production-ready data services with correctness, auditing, and operational reliability in mind.