Sovereign Data for AI with Python
summary
Claims a full Pythonic sovereign AI stack that runs on air gapped or hybrid infrastructure and uses open source pieces: PyPI and Conda library server, a container registry, S3 storage and a database with a vector index. Demonstrates a data layer that decouples feature training and inference, registers external tables (BigQuery, S3), adds residency tags and policy rules, launches prepackaged Jupyter environments, saves models optionally, and exposes a serverless online DB with capped queries while noting a live demo hiccup and that free deploys omit compute.
AI-written summary · about
chapters (11)
- 0:00Sovereign data for AI
- 1:06Sovereign data and AI
- 2:11Global data sovereignty trends
- 4:20AI infrastructure gaps
- 6:27Python in sovereign AI
- 8:11Requirements for sovereign AI
- 10:20Architecture for sovereign AI
- 12:47Decoupling data in Python
- 14:10Live demo: sovereign platform
- 20:25Operationalizing sovereign AI
- 22:52Summary
related talks
30mCreating an AI-Python Infrastructure that works with a reasonable scale
33mAI sovereignty, provenance and compliance
15mEnterprise AI Runs on Trusted Data
24mBuilding Sovereignty: Where Your AI Runs Matters
42mAI – Addressing the Challenges of Today’s Data Scientists
49mScaling Data in a Sovereign AI Platform
16mKeynote: Accelerating AI/ML Development with Open Source Software
01h 03mDo Python and Data Science Matter in Our AI Future?