


Zero to Pipeline: When Data Connectors Just Work
EuroPython 2026 · Krakow, Poland
Data engineers spend hours writing the same connector boilerplate (auth flows, pagination loops, retry logic) before a single byte of real data moves. I built zero-to-pipeline to fix that: one CLI command, and the LLM discovers endpoints, infers the auth type, and writes the config. At EuroPython 2026 (Europe's largest Python conference, 25th anniversary edition in Krakow) I ran a live fraud-detection demo: 15,000 synthetic transactions, Feast feature views, MLflow, Prometheus, and a Linear issue, orchestrated as a DAG and done in 4.4 seconds. The talk collapses days of connector boilerplate into a single shell command.