Description
In a rapidly evolving technological age where large language models such as OpenAI’s GPT series are becoming increasingly relevant, software developers and architects are faced with the fascinating task of using these advanced technologies to optimize chat applications. The key to high-quality interactions with these AI models is the art of designing requests precisely and thoughtfully.
This full-day workshop dives deep into the art of prompt engineering. You will learn the basics of large language models such as ChatGPT and, building on this, you will learn important prompt engineering tricks.
This workshop is specially tailored to the needs of software developers in the architecture role and software architects.
The icing on the cake of the workshop is building your own chat frontend that gets the most out of the OpenAI API. Among other things, we will give the LLM the opportunity to correct itself.
Note: To participate in this training, you need a ChatGPT+ account.
Agenda
-
Definition of terms, introduction to LLMs
-
Prompt engineering techniques with ChatGPT 4
-
Diagrams and Python Code Sandbox
-
Limits of ChatGPT
-
Pushing the boundaries with OpenAI API and your own chat frontend
-
Creating your own sandbox using Java as an example
Your Benefits
You will learn about the background of LLM chats such as ChatGPT.
This will give you a better understanding of the possibilities and limitations.
You will deepen your knowledge of various prompt techniques.
You will learn how to use LLM chats profitably for architecture work.
You will see the OpenAI API in action.
Audience
The workshop is aimed at developers, software architects and technology experts who are considering the use of LLM in chat applications or have already gained initial experience. It shows what is possible with this technology and prepares participants for its use.
Training Objectives
How does a large language model work?
What do you need to be aware of when using it and what are the limitations?
Beneficial use of ChatGPT and other large language models
How can the limitations of the ChatGPT frontend be softened?
How can we successfully use LLMs in software architecture?
Your Trainers
Ralf D. Müller
Docs-as-Code, Softwaredokumentation
- Documentation-as-Code
- Fantastic diagrams
- Prompt engineering for architects and developers
Ralf hat in über 25 Jahren in der Softwarebranche seine Fähigkeiten kontinuierlich verbessert und sich vom Softwareentwickler zum erfahrenen Software-Architekten entwickelt. Neben dem Design von Softwarearchitekturen legt er großen Wert auf die klare Kommunikation und Dokumentation seiner Ideen. Mit diesem Ziel vor Augen hat Ralf das Open-Source-Projekt docToolchain ins Leben gerufen, das sich mit der effektiven Dokumentation von Softwarearchitekturen beschäftigt. Er ist auch verantwortlich für die Erstellung der verschiedenen Formate des arc42 Templates. Als CPSA-Advanced-zertifizierter Softwarearchitekt teilt Ralf sein Wissen und seine Erfahrung in Schulungen und Vorträgen, um anderen zu zeigen, wie sie Architekturdokumentation einfach und aktuell halten können. Video: Mehr als Pfeile und Kästchen – Architekturdiagramme im Softwarearchitektur-Stream.
Technical Information and Books
Talk: Using AI in Software Design: How ChatGPT Can Help With Creating a Solution Architecture
In this talk, Ralf D. Müller looks at the integration of AI into software design and in particular introduces ChatGPT, a Large Language Model (LLM). Watch this talk if you are interested in using ChatGPT in software architecture. Click here for the talk.