Uploaded August 2026 | Updated September 2026, 2 weeks ago
[EuroPython 2026 - S4 on 2026-07-15]
π€ *Zero to Pipeline: When Data Connectors Just Work by Olusola Akinsulere*
π https://ep2026.europython.eu/session/zero-to-pipeline-when-data-connectors-just-work
π Abstract:
Remember spending entire afternoons configuring data connectors, debugging OAuth flows, and writing custom parsers just to ingest data from a single API? What if you could just say pipeline source add feast and have your pipeline automatically discover the API, handle authentication, and start ingesting data no config files, no setup wizards, no SDK imports?
In this 30-minute talk, I'll show you the reality behind self-configuring data pipelines that sound too good to be true but actually work. Using a fraud detection use case Feast feature store, MLflow experiment tracking, Prometheus metrics, and Linear for issue tracking we'll explore how intelligent connectors leverage LLM-driven API discovery, automatic credential management, self-healing auth, and cursor-based checkpointing to eliminate the traditional pain points of data ingestion.
We'll be honest about the tradeoffs too.
---
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License: creativecommons.org/licenses/by-nc-sa/4.0
[EuroPython 2026 - S4 on 2026-07-15]
π€ *Zero to Pipeline: When Data Connectors Just Work by Olusola Akinsulere*
π https://ep2026.europython.eu/session/zero-to-pipeline-when-data-connectors-just-work
π Abstract:
Remember spending entire afternoons configuring data connectors, debugging OAuth flows, and writing custom parsers just to ingest data from a single API? What if you could just say pipeline source add feast and have your pipeline automatically discover the API, handle authentication, and start ingesting data no config files, no setup wizards, no SDK imports?
In this 30-minute talk, I'll show you the reality behind self-configuring data pipelines that sound too good to be true but actually work. Using a fraud detection use case Feast feature store, MLflow experiment tracking, Prometheus metrics, and Linear for issue tracking we'll explore how intelligent connectors leverage LLM-driven API discovery, automatic credential management, self-healing auth, and cursor-based checkpointing to eliminate the traditional pain points of data ingestion.
We'll be honest about the tradeoffs too.
---
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License: creativecommons.org/licenses/by-nc-sa/4.0
