Workshops and courses

ELIXIR node event de.KCD Summer School 2024: Deploying LLMs in the Cloud

View event

Date: 16 - 20 September 2024

Timezone: Berlin

Duration: Approx. 4h per day via Teams, including breaks. Starting and ending time are provisional and subject to change.

Loading map...

Ever wanted to look behind the – technical and organizational – scenes of large language models ? Already have your own AI research project sketched out but unsure of how to actually get started? Looking for cloud storage to power up your LLM but getting caught up in regulatory requirements?

This summer school will provide practical answers by working with you on your own hands-on sample LLM in the de.NBI cloud. Instead of struggling through the steps on your own and reinventing the AI wheel over and over again, we at the de.KCD project aim to pool our joint expertise and provide you with a shared, tried-and-tested workflow for cutting-edge research.

Over the course of the week, we will walk you through the entire lifecycle of LLM deployment in the cloud. We will take a deep dive into the individual stages of the “day zero one two operation” concept of software engineering – where day 0 is the planning and design phase (including setting up your cloud access and drafting your data management plan), day 1 is go-live (installing and deploying your LLM), and day 2 is continuous operations (including backups and user authentication) and everything else that comes after going live.

Contact: Please email Lisa Schaak: [email protected]

Keywords: Cloud Computing, hands-on, LLM, Summer School, Research Data Management, Data Protection, generative AI, online, Machine Learning, de.KCD

City: Berlin

Country: Germany

Prerequisites:

  • Basic terminal knowledge
  • Some experience with programming languages, ideally python, but R, perl, SQL etc. are also fine
  • Basic knowledge of docker

Learning objectives:

Participants will deploy their own LLM in the de.NBI cloud via a step-by-step guided journey throughout an entire lifecycle from design via deployment to productive operation.

  • Day 1: How to Cloud – Introduction, de.NBI Cloud & Project Idea
  • Day 2: How to Data – Data Management, Data Protection & Data Ethics
  • Day 3: How to LLM – Installation & Deployment of a Large Language Model
  • Day 4: How to Service – Ensuring Continuous Support throughout Operations
  • Day 5: How to Success – Real-Life Examples, Q&A and Discussion

Organizer: de.KCD & BIH @ Charité

Eligibility:

  • First come first served

Target audience: Everyone interested in digging into the technical and organizational details of deploying large language models for research via cloud infrastructure. Most of our trainers come from a biomedical and IT background. However, we are looking forward to an interdisciplinary exchange with research-developers and developer-researchers from all kinds of academic disciplines.

Capacity: 40

Event types:

  • Workshops and courses

Tech requirements:

  • Internet connection
  • Linux terminal / shell
  • Browser
  • Microsoft Teams (or browser-bound version)
  • Camera, microphone

Credit / Recognition: We will not be able to offer any ECTS credit points but we can provide you with a certificate of attendance upon request.

Cost basis: Free to all

Activity log