Unstructured Text Analysis With BYO-LLM and NVIDIA GPUs on VantageCloud
Organizations are under increasing pressure to quickly deploy valuable generative AI applications and demonstrate a strong return on investment (ROI). However, deploying gen AI comes with significant challenges such as selecting the right large language model (LLM), while ensuring data security, privacy, and cost control.
The Bring Your Own LLM (BYO-LLM) capability within Teradata VantageCloud offers a streamlined approach to experimentation, deployment, and performing inference using LLMs directly within the database. This capability reduces complexity and costs, while ensuring that developers and data professionals retain complete control over their data in a secure, flexible and trusted environment.
Join our live demo to see how to bring Hugging Face open-source language models into Teradata VantageCloud and supercharge unstructured text analysis with GPU acceleration — right where your data already lives.
We will explore:
- Setting up VantageCloud: Configuring analytic GPU compute groups and connecting using teradataml
- Model deployment and inference: Step-by-step guide on downloading and deploying the 'tner/roberta-large-ontonotes5' model from Hugging Face and conducting batch inference jobs for entity recognition
- Extracting insights from unstructured data: Demonstrating how to analyze texts, emails, PDFs, customer communications, call center transcripts, and other forms of unstructured data to extract valuable insights
Speakers include:

Developer Advocate at Teradata
Janeth Graziani enjoys leveraging her expertise in technical content creation to help developers navigate and incorporate new tools into their tech stacks.

Product Management Director at Teradata
Jefferson Uy is a ClearScape Analytics™ Product Manager at Teradata, focusing in open source analytics solutions that help businesses leverage their data for strategic insights. He has a passion for world history and enjoys watching video documentaries of them whenever he can.