Tom Kaltofen builds open-source data tools. He created mloda, a Python framework that untangles transformation logic from data, allowing developers to reuse feature engineering across Pandas, Spark, and DuckDB.
As the CTO of mloda, his work targets the core reason AI agents hallucinate in production: unreliable data context. He focuses on designing deterministic context layers that decouple compute logic from data extraction, ensuring agents receive clean, validated information without brittle pipelines.
Tom lives in Berlin, where he actively maintains the mloda project on GitHub and PyPI.