The InfoQ Podcast

Sam Partee on Retrieval Augmented Generation (RAG)

10 snips
Jan 29, 2024
Guest
Sam Partee
Sam Partee, principal engineer at Redis, discusses Redis' vector database offering, different approaches to embeddings, enhancing language models with search components, and the use of hybrid search in Redis. They also explore the potential applications of retrieval augmented generation (RAG) technology and the challenges of running large language models on-prem.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
ANECDOTE

Sam Partee's Redis Vector Journey

  • Sam Partee shared his journey integrating Redis with vector databases and embedding frameworks like Langchain.
  • He learned many best practices through direct customer work and open source development.
ADVICE

Choose Vector DB by Use Case

  • Choose a vector database based on your use case, like Redis for real-time and dynamic vectors.
  • For static large datasets, cheaper solutions like FAISS on S3 may be more practical.
ADVICE

Best Use Cases for Redis Vectors

  • Redis vectors excel for chat conversation memory, semantic caching, and live recommendation systems with fast latency needs.
  • Avoid Redis vector DB for static, low QPS use cases as it can be cost-ineffective.
Get the Snipd Podcast app to discover more snips from this episode
Get the app