Prompt Engineering: MasterClass Prague Data Ethics Lab & PwC & Omnicrane

Event Description:

Through live demonstrations of multiple Prompt Engineers with different online experiences (University, PwC, Startup) demonstrate to participants in an iterative and incremental way what can be gained by working with LLM. To inspire imagination, skills and stimulate discussion among participants about the best practices in the area of Generative AI & Prompt engineering.

You will get an interactive and iterative explanation according to the procedure: From LLM basic questioning to solving real and advanced business problems

Speakers:

• Richard A. Novák, CEO Omnicrane, Chairman Data Ethics Lab
• Michal Caninec, VC Analyst at V# Venture Studio
• Kamil Matula, Data Researcher at Silesian University
• Egor Suslin, Data Scientist at PwC
• Jiří Korčák, Prague Data Ethics Lab

Structure of live workshop:

1.      Intro to LLMs and their limits & hyperparameters of GPTs, Context vs tokens window, Defining purpose & generated artifacts as text, image, code, video…etc., API & Plugins, Pricing, Performance vs Size of models, Best practices for generation of artifacts as text, image, code, video…etc., Evaluation of Pricing vs Performance vs Size of models

2.      Naming typical use cases as Questions Answering, Language translation, Sentiment analysis, Summarization, Code generation, Data Analytics, Chatbot response, Classifications, Chaining of models…etc.

3.      Experiments using different modifiers, styles, perspectives, digital personas and formatting, From zero-shot to few-shot inference, using data as RAG method, Iterations & fine-tuning outputs, How to measure performance of Prompts  

AGENDA:

14:00 – 14:15  Introduction of „Prompt Engineers“,

14:15 – 15:00  Overview of principles & methods of LLMs

15:00 – 15:30  GenAI artifacts as text, image, video generation...

15:30 – 16:00  Data analytics & using LLM

16:00 – 16:30  Complex Business problem solving via LLM & wflows

16:30 – 17:00  Discussion & Networking

Event Format:  English speaking, Offline + YouTube recording

Organizer: Prague Data Ethics Lab