Which to Use: MCP, Function Calling, or Plugins
Three AI integration patterns -- MCP, function calling, and plugins -- each fits a different shape. Here is the decision tree I use with clients.
Read more
Labyrinth Analytics
Agentic Workflow Consulting & Data Engineering
Illuminating the dark corners of your data.
Data engineering, warehouse architecture, and agentic AI workflows for organizations ready to find their way forward.
What We Do
Every dataset tells a story. Most of it is buried in the dark. We help you find it.
Data Pipeline Design
Your data sources are scattered and unreliable. We build pipelines that bring order to the chaos -- clean, dependable flows you can trust at 2 AM.
ETL & Data Warehouse
The vault where your data lives matters. We design warehouses built on dimensional modeling, incremental loads, and architecture that scales with your questions.
AI Workflow Design
From intelligent analytics to autonomous agent systems. We design AI workflows that range from LLM-powered validation to fully scheduled agentic pipelines with human review gates -- built to do the work, not just describe it.
Recent Thinking
Notes from the field on agentic workflows, data engineering, and the tools we build along the way.
Data Quality Lessons for AI Agent Pipelines
The same data quality failures that plagued ETL pipelines in the nineties are resurfacing in AI agent workflows -- and the fixes are identical.
Read moreLoreDocs
LoreDocs: a local knowledge vault
LoreDocs is a local knowledge vault for data engineers and AI practitioners -- durable, searchable storage for artifacts that outlive a single chat.
Read moreOur memory layer for Claude just launched
Coming Soon
A lifetime deal on the Lore family
LoreConvo and LoreDocs, yours for good. Join the waitlist to be first through the door.
See the Lifetime DealReady to find your way through the labyrinth?
Whether you are untangling legacy systems or building from scratch, we would love to hear about the challenge.