Leveraging LLMs to Build Intuitive AI Experiences With JavaScript

The presenters are a Developer Relations Engineer from Couchbase and Roy, who works at IBM.

The workshop discusses technologies such as LLMs, generative AI, VectorSearch, TensorFlow.js, Transformers.js, and various JavaScript libraries.

Vector databases store vector embeddings of data, which can be used for semantic search and retrieval augmented generation (RAG) to make AI models context-aware.

AI applications can be made production-ready using local models by deploying them in containerized environments like Docker or Kubernetes, or on virtual machines, ensuring privacy and control over data.

Recommended models for generating JavaScript code include Code Llama, Dolphin Mistral, and Phi-3, which are specialized and lightweight for code generation tasks.

The live workshop focuses on leveraging large language models (LLMs) and generative AI with the help of JavaScript for front-end development.

Yes, TensorFlow.js can run AI models directly in the browser, offering benefits like lower latency, higher privacy, and reduced server costs.

OLAMA is a tool that allows you to run open-source large language models (LLMs) locally. It supports models like LLAMA3, Dolphin Mistral, and others, which can be served locally for offline use.

Using local models offers better privacy since data isn't sent to external servers, reduced latency, and potentially lower costs by avoiding cloud service fees.

Few-Shot Prompting involves providing an LLM with a few examples of questions and their desired formats to guide the model in generating more accurate responses.