Building Multimodal Agents With NVIDIA Nemotron and RAPIDS - Part Two

Turn multimodal document and image descriptions into GPU-accelerated topic models. Use BERTopic, cuML UMAP, and cuML HDBSCAN to cluster and inspect text collections extracted by Nemotron in Workshop 1.

In this workshop you will learn how to:

1. Use RAPIDS cuML UMAP and HDBSCAN to accelerate topic modeling stages on GPU
2. Convert OCR outputs and visual descriptions into analysis-ready topic records
3. Design a LangGraph agent for document ingestion, topic modeling, and summarization

Prerequisites:
Python 3 fundamentals, exploratory data analysis workflows, and basic microservice concepts. It is also recommended to attend Part 1 of this workshop series.

Recommended Resources:
NVIDIA NIMs intro
NVIDIA RAPIDS intro
LangChain Academy's Intro to LangGraph