Lead product and architecture for AI-powered tax research tools at EY Foundry — EY’s innovation hub focused on preventing technological stagnation in the global tax practice.
•
Designed and built multi-agent LLM systems that deconstruct complex legal/tax questions into structured reasoning chains, retrieve targeted RAG results, and synthesize expert-grade memo outputs across domains like sales and use tax.
•
Architect end-to-end systems that integrate Azure AI Search, Data Factory, and Machine Learning services to deliver scalable, secure semantic search and retrieval workflows across massive legal corpora.
•
Developed prompt-routing frameworks that dynamically map user queries to the correct content type (statutes, case law, regulations, or news) using vector embeddings, metadata filtering, and few-shot prompting.
•
Built feedback loops, synthesis orchestration logic, and internal tools to generate and reason over knowledge graphs, enabling recursive improvement and deeper understanding of tax guidance.
•
Own core product responsibilities: write user stories, manage backlogs, lead standups, and drive alignment across engineering, design, and business stakeholders.
•
Lead an engineering team focused on system implementation and integration, while personally owning high-level architecture, LLM prompt design, and complex logic development
