Designed and built the data foundation for a health coaching company's AI assistant system enabling $100K+/mo additional client revenue. Architected a Supabase schema with row-level security, built a client identity resolution map unifying data across Google Drive, Discord, multipe CRM databases, and spreadsheets, and selected and benchmarked embedding models against real client data (chat logs, call transcripts, YouTube content). Delivered the first ingestion pipeline in n8n, an architecture playbook for remaining data sources, and a full Zapier migration assessment for ~50 workflows. This is the foundation layer that the coach briefing tool, tiered chatbots, and analytics dashboard will build on.