Practically Intelligent Vector Databases, Embeddings, and a history of Deep Learning with Leo Dirac
32 snips
Jun 8, 2023 Former Engineering Lead behind Deep Learning at AWS, Leo Dirac, shares a walk through history and key takeaways for builders in the AI/ML space. They discuss the importance of vector databases, comparing different options, and the challenges of computer vision. Leo also talks about his new venture, Groundlight.AI, and its role in simplifying computer vision for engineering leaders.
AI Snips
Chapters
Transcript
Episode notes
Choose DB By Cost, Latency, And UX
- Evaluate vector databases by cost, read/write latency, and storage efficiency.
- Also consider logical features like namespaces, access controls, and developer tooling for real-world operations.
Developer-First Tools Win Adoption
- Prefer vector DBs with simple SDKs, instant sign-up, and Python wrappers for quick prototyping.
- Choose developer-first products to test ideas before committing to complex self-hosted stacks.
Dynamic Few-Shot Retrieval Boosts LLMs
- Dynamically retrieving similar examples from a vector DB as few-shot context improves LLM accuracy.
- Even simple arithmetic tasks benefitted when GPT was shown related solved examples.