- Architected a multi-tenant, AI-native SaaS platform automating end-to-end operational workflows for service businesses, replacing CRUD UIs with agentic conversation flows.
- Engineered an autonomous AI Receptionist & Intake engine (Python, LangChain, LangGraph) orchestrating Voice, SMS, Web, and Social channels with zero human intervention.
- Built a localized computer-vision pipeline (PyTorch, OpenCV, vision LLMs) that extracts structural data points to feed a dynamic pricing model.
- Implemented sub-graph Agentic RAG (Pinecone) turning voice dictations and service history into actionable planners, upsells, and retention models.
- Designed predictive modeling (scikit-learn) forecasting cancellations, no-shows, and technician utilization; generated automated mitigation workflows.
- Built a natural-language BI layer over SQL/NoSQL schemas for non-technical operators.
PythonTypeScriptLangChainLangGraphPyTorchOpenCVPineconeDockerAWSPostgreSQLRedis