SuperDuperDB: Bring AI to your favourite database!
summary
SuperDuperDB claims to run models where the data lives so you never copy or ETL data, offering a Python package plus a deployment orchestrator using Ray, MongoDB connectors, an artifact store, and Apache 2.0 open source code. Demo builds a custom OpenCV video frame extractor, vectorizes frames with CLIP, wires a vector index with listener kept up to date via CDC, and returns exact timestamps for “elephant” hits — a working self‑hosted multimodal search. no data movement
AI-written summary · about
chapters (25)
- 0:00intro
- 0:20preamble
- 0:57data and ai live in separate silos
- 1:24bringing machine learning to production is very complicated
- 1:59ai needs and vector search need to be integrated directly with the data source to remove overhead
- 2:30superduperdb brings ai to the database
- 2:51building, deploying and managing ai applications without moving data
- 3:17superduperdb: the centerpiece of the modern data-centric ai stack
- 4:13all-in-one platform for all data-centric enterprise ai use-cases
- 4:43python-first: build for developers with the ecosystem in mind
- 5:05open-source: free apache 2.0 software
- 5:24how superduperdb works
- 8:23connect
- 8:43code to query
- 9:25code to create custom datatype
- 10:09code to create inline model
- 11:52create custom superduperdb model
- 12:36apply model
- 13:04apply model to data updates in real time
- 13:43create vector-index
- 14:07create superduperdb stack
- 14:28parametrize superduperdb stack
- 15:09share superduperdb app
- 15:25setup even the most complex ai workflows effortlessly
- 21:56learn more
related talks
39mPython and SQL: Better Together, Powered by DuckDB
45mManaging Data & Analytics Projects Effectively
25mProcessing Biggish Data with DuckDB and Python
25mA quack of all trades: DuckDB, a versatile analytical database to keep in your toolkit
40mData Intelligence Platform on AWS
34mData Wrangling with DuckDB
37mAccelerating Python Analytics by In-Database Processing
30mDuckDB: Bringing analytical SQL directly to your Python shell
