datumagic
distributed systems • data lakehouse • physical AI
Deep dives into data infrastructure at scale — open-source table formats, multimodal data for AI training, and the internals of modern data systems.
Latest blogs

Spinning the Data Flywheel: From Curation to Fleet (opens in a new tab)
The stages that turn labeled data into a model on the road—curation, scenarios, data loading, training, and evaluation—and the challenges each raises at scale.
Physical AIModel TrainingLanceOctober 28, 2026appliedintuition.com
Apache Hudi 2025: A Year In Review (opens in a new tab)
A look back at Apache Hudi in 2025: releases, community growth and ecosystem milestones.
Apache HudiDecember 29, 2025hudi.apache.org
Maximizing Throughput with Apache Hudi NBCC: Stop Retrying, Start Scaling (opens in a new tab)
Learn how Hudi's Non-Blocking Concurrency Control eliminates retry storms for concurrent writers, maximizing throughput in streaming and mixed workloads.
Apache HudiConcurrency ControlStream ProcessingDecember 16, 2025hudi.apache.org
Talks

Building the Autonomous Driving Data Flywheel (opens in a new tab)
Reverie Summit 2026 (panel)


