Workshops Are New In 2024

Generative AI is easy!

Look at ChatGPT, you ask a question, you get an answer! How difficult could it be?

This workshop is tailored for those seeking to understand the application of Generative AI within enterprise environments, focusing on the requirements for security, scalability, and extensibility. The workshop will delve into various topics and include demonstrations throughout:

1. An exploration of Large Language Models (LLMs), discussing their purpose, how they are constructed, and how they are made available for use.
2. Detailed insights into Prompt Engineering and how it enhances the relevance and coherence of AI chatbot responses in real-time conversations.
3. Demonstrations of Zero-Shot and Few-Shot Prompts to illustrate their effectiveness with OpenAI's LLMs.
4. A discussion on the role of Orchestrators, including Semantic Kernel (SK) from Microsoft for integrating LLMs with conventional programming languages, and LangChain for application development with LLMs.
5. A deep dive into Retrieval Augmentation processes, including RALM, RAG, ReAct, and COT patterns, to enrich LLM responses with company data.
6. An introduction to Embedding Models and Vectorization, explaining the transformation of high-dimensional data into low-dimensional vectors for easier machine learning processing.
7. A look at Vectorization Databases, which store data as mathematical vector embeddings, and a discussion on various available platforms like Azure AI Search, Pinecone, Chroma, Faiss, and Azure Postgresql.
8. An overview of security practices, including Red Team exercises and Security Hardening, aimed at identifying and mitigating vulnerabilities.
9. An explanation of Row level Security (RLS) for limiting access to data within a database, detailing its application in securing Generative AI environments.

Participants will gain knowledge on operating LLMs and orchestrators, managing custom data, and implementing security measures to ensure the safety and privacy of organizational data while maintaining scalability and extensibility.

Limited to 15 attendees.