LLM has revolutionized the use of AI. But how do you create and train an LLM? And if you use an existing, which LLM’s exists and how can you make good use of an LLM (technically and resposibly)?
This course is for you who wants to understand LLM’s “underneith the hood”, and for you who are in the process of making good use of a LLM by fine tuning it and putting it to use in your business.
This is a one day course, with hands-on coding.
Building solutions using Large Language Models
Course Agenda
Foundations
- What is a LLM? Definitions
- History of LLM development
- LLM Design
- Prompting
Creating LLM’s
- The Neural Architecture (for example Transfomer models)
- Model selection, size and scaling
- The training process
- How to overcome limitations (for example computing power)
- Fine tuning of LLM’s with own data
- RAG (retrieval augmented generation)
- LLM use cases (with examples)
Ethics and privacy
- Dealing with prejudice during training
- Ethics & Privacy
Closing
- Limitations of LLM’s
- Future trends for development and research
Course prerequisites
- Foundational ML and AI knowledge, equal the skills from “AI for developers”
- You will also need a ChatGPT4 account (typically costs $20)
- Bring a laptop!
Course details
Course details
- Price: 9000 SEK (all prices ex. VAT)
- Early Bird: 6000 SEK (only first five signups, latest 2 month before course starts)
- Place: Convendum, Central Stockholm (onsite)
- Date -> check bottom of the course page!
About the trainers
Birger Moëll
Birger is a PhD student in machine learning @ KTH with years of experience explaining AI.
“I am passionate about teaching people hands-on tips to get started working as AI developers, and am excited about the opportunity to teach this course to empower individuals with the necessary skills and knowledge to succeed in this exciting field
Fredrik Fridborn
Fredrik has worked as engineer at AI startups, holds a MSc in electrical engineering and has spent years making AI understandable to those without a degree in mathematics.
Who the course is for
- Developers
- Data scientists
- CTO’s