Kevin Lin

Contributed to the development and launch of ASK B, Bloomberg's agnostic AI conversational interface for the Terminal, enabling finance professionals to ask complex natural-language questions and receive synthesized insights from news, research, documents, market data, and analytics. Designed and implemented core components of a multi-agent orchestration system, enabling specialized AI agents to retrieve, analyze, and generate contextual answers with source attribution and auto-generated code for seamless extension in Excel or BI tools. Built reusable ASK B workflows to standardize multi-step research tasks, speeding up terminal users' workflows. Integrated enterprise-grade guards and retrieval-augmented generation, ensuring reliability and low hallucination in a regulated financial environment. Supported the 2026 roadmap by enabling unified data access with firm-proprietary knowledge, including document upload/analysis, alternative data synthesis, and shareable workflows to augment th

Scaled the Key Stone real-time stream processing platform using Apache Kafka, Apache Flink, and Apache Spark to process 2+ trillion events and 10+ petabytes of data daily, powering personalization engines that drive ~80% of Netflix viewing across 190+ countries. Designed and enhanced self-service streaming processing capabilities used by thousands of internal use cases, including recommendation engines, A/B testing, fraud detection, and operational analytics. Contributed to core personalization and recommendation architecture by incorporating real-time user behavior signals to improve viewer retention and engagement at massive global scale. Supported live-streaming infrastructure initiatives enabling low-latency delivery and dynamic real-time recommendations for major live events. Built and scaled Streams as a Service (SPaaS) on top of Key Stone, enabling hundreds of teams to build and operate custom real-time applications. Designed real-time recommendation update mechanisms for live e