Our Blog
Insights, tutorials, and case studies on AI, data science, and churn prevention
How We Cut Agent Costs 10× by Turning Traces into Tools
Our sales agent ran 11× more expensive on a frontier model, for half the work. So we used it as a teacher instead: its traces became deterministic tools, playbooks, and verification gates a cheap model can follow.
Read More →
MetAgent Local — Running a Multi-Agent Analytics System on Your Own Infrastructure
How to run the full MetAgent multi-agent analytics system on your own infrastructure with Docker Compose. No cloud accounts, no data leaving your network — full LLM observability included.
Read More →
Building an AI-Powered SQL Assistant with Metabase
How I built a two-agent system inside Metabase that lets anyone query a database in plain English — no SQL required. The first step toward MetAgent.
Read More →
AIR: K-Nearest Neighbors
I used KNN to figure out the optimal nightly price for our Airbnb in Medellín. A practical walkthrough of scraping, modeling, and making a real investment decision with data.
Read More →
Optimizing our Delivery Process with Data Science
At Mercadoni we were underpaying shoppers 40% of the time. Here's how quantile regression fixed it — cutting transaction costs 70% and manual transfers by 80%.
Read More →