Postgres-powered AI: Running an End-to-End AI Platform with Postgres on Azure
summary
An end-to-end AI platform built on Postgres uses Postgres JSON fields for heterogeneous metadata, leverages ACID concurrency to avoid bespoke locking, and integrates Airbyte, JupyterHub, Superset/Plotly Dash, Kubernetes, Helm, Terraform and Python for ingestion, notebooks, dashboards and orchestration. Migration from an AWS-first SaaS to a BYOC Azure deployment completes in under 3 months, requiring only cloud service remaps (S3→Blob, RDS→Azure SQL) while core model metadata, pipelines and runtimes remain unchanged and fully functional.
AI-written summary · about
chapters (8)
- 0:00Intro
- 4:03Data Ingestion with Postgres
- 6:24Data Processing with Postgres
- 7:32AI/ML Workloads with Postgres
- 9:22Data Visualization with Postgres
- 11:29An "Enterprise" AI Challenge
- 19:53Azure Compatibility Journey
- 23:07Summary
related talks
44mPostgres and OpenAI - a match made in Azure
46mTransform applications with AI and Azure Database for PostgreSQL
12mThe mechanics of integrating AI and vector search into Azure Database for PostgreSQL
34mData modeling for those who want to save time and money!
57mDeveloping powerful and intelligent apps using the Azure managed PostgreSQL service
08mKeynote
56mSeamless Migration to Azure Database for PostgreSQL
35mPostgres in Artificial Intelligence