Generative AI

This section covers:

• “Agents and Tools for Generative AI” — guides on building LLM-powered agents (e.g., Code Agents, Visual Agents), creating custom tools, and using the LLM Mesh.

• Architecting GenAI-powered components in Dataiku (agents, tools, workflows)

• Integrating LLMs with external systems and tooling

• Building robust, tool-enabled AI workflows using Dataiku’s GenAI ecosystem.

• Other core GenAI topics include retrieval-augmented generation (RAG), prompt design, monitoring, inline tools, etc.

Agents and Tools

• Creating an Inline Python Tool

• Creating a custom tool

• Leveraging a custom tool in a Visual Agent

• Creating and using a Code Agent

• Creating a custom agent

• Multi agents: sequential workflow

• Building and using an agent with Dataiku’s LLM Mesh and Langchain

• Defining and using tools with the LLM Mesh

• Creating an LLM-based agent that uses multiple tools

• Building a Web Application with the agent

• Building Auto Prompt Strategies with DSPy in Dataiku

• Using the LLM Mesh to parse and output JSON objects

• Adding traces to your Agent

• Building your MCP Server in Dataiku

Multi-modal capabilities

Images and text

• Image generation using the LLM Mesh

• Mixing text and images: Images labeling with the LLM Mesh

Natural Language Processing & Generation

• Programmatic RAG with Dataiku’s LLM Mesh and Langchain

• RAG: Improving your Knowledge Bank retrieval

• Creating and using a Knowledge Bank

• Using LLM Mesh to benchmark zero-shot classification models

• Zero-shot text classification with the LLM Mesh

• Few-shot classification with the LLM Mesh

• Using standard OpenAI API calls via the LLM Mesh

• Customizing a Text Embedding Model for RAG Applications

Webapps

Headless API

• API Endpoint on LLM

Dash

• GPT-powered web app assistant

• LLM based agent

Gradio

• First application

• LLM based agent

Voilà

• First application

• LLM based agent