From Theory to Terminal: A Practical Guide to Running LLMs on GCP
On September 30th we are again bringing you the insight you look for about the building blocks of Large Language Models (LLMs). Join us online to learn more about using offline models on GCP! ☁️
Attendance is free of charge but registration is required
💡 Abstract
Moving beyond public APIs, this session explores the practicalities of running powerful open-source Large Language Models (LLMs) within your own Google Cloud environment. We will demystify the business case for a private AI strategy, focusing on how hosting models in your GCP project enhances data security, provides predictable costs, and grants ultimate control. From selecting the right Compute Engine instances to leveraging the experimentation tools in Vertex AI, you will gain a clear roadmap for efficient deployment. We will then demonstrate how to programmatically interact with these models and use Retrieval-Augmented Generation (RAG) to securely query your internal documents.
Attendees will leave with the confidence to deploy their first private, customized LLM using Google Cloud's powerful infrastructure.
👨💻 Suggested audience: Architects and IT experts who are not specialized on the AI technologies, but they are aware of the basic terminology of AI, and are interested in the technology of AI. We are building on the knowledge we shared at the first episode of the series Technology behind the Magic of AI: Introduction, however we repeat the most important elements of the first episode as a refresher.
🗣️ Language: English
✏️ Save the date, and don’t miss out on this opportunity to learn, share, and innovate together!
------------------------
📌 Detailed Agenda:
1. Recap of the previous episode & explaining why we should consider offline models
2. Introduction to the Offline LLM Ecosystem, the sources and the most important file formats to use
3. Hardware requirements of the Offline LLM models
4. How to run and program your LLM: Demos (installing models on Ollama, developing AI based applications in Python)
5. How to make models smarter (Training with own data)
6. Conclusion and Q&A
------------------------
SpeakerBalázs Molnár - Deutsche Telekom IT Solutions (Cloud Architect)
Balázs is a Cloud Architect with more than 30 years of experience in Enterprise Software and Hardware technologies, including 8 years of Cloud technology including Cloud infrastructure and AI. He spent majority of his career at International IT companies like Oracle and Apple. Balázs has 15 years of experience in International Enterprise projects, customer advisory and presales activities in topics like Software implementation, application environment architecture and sizing, Business Case development and Enterprise Architecture. Today he’s working as a Cloud Architect at the Google Team of Cloud Professional Services of Deutsche Telekom IT Solutions.
------------------------
Data privacy information: Please note that the event will be recorded by DT-ITS and the recording may include the audience's activity during the Q&A sessions as well. The recording will be published on DT-ITS’s YouTube channel: Clouders Club, and on DTAG internal communication platforms to help professional development, share information and knowledge. By actively taking part in the Q&A session we assume that you consent to the recording of your voice during the event. For any request regarding the described data processing please contact us on this email: HU_DT_TSI_CS_BO@t-systems.com
Agenda
---
Speaker
Balázs Molnár - Deutsche Telekom IT Solutions (Cloud Architect)
Balázs is a Cloud Architect with more than 30 years of experience in Enterprise Software and Hardware technologies, including 8 years of Cloud technology including Cloud infrastructure and AI. He spent majority of his career at International IT companies like Oracle and Apple. Balázs has 15 years of experience in International Enterprise projects, customer advisory and presales activities in t…
Host
Kamilla Kerschner - Deutsche Telekom IT Solutions (GDG Organizer)
Organizer of GDG Cloud Budapest community
Complete your event RSVP here: https://gdg.community.dev/events/details/google-gdg-cloud-budapest-presents-technology-behind-the-magic-of-ai-part-2-using-offline-models-on-gcp/.