Building GenAI-Powered Applications - A Workshop for Software Engineers

Join a comprehensive 3-hour workshop led by industry experts Stefan Krawczyk (DAGWorks CEO) and Hugo Bowne-Anderson (Independent Data Scientist) to master the fundamentals of building generative AI-powered applications. Learn essential skills through first principles thinking, covering prompt engineering, monitoring, evaluations, and handling non-determinism in AI systems. Explore practical applications like PDF querying while gaining insights into engineering challenges specific to AI development. Through hands-on sessions, discover how to generate and refine emails using Language Models (LMs), implement human-in-the-loop workflows, and extract structured data from PDFs. Master crucial production aspects including observability, tracing, error handling, and Retrieval Augmented Generation (RAG) components. Dive into prompt optimization, synthetic data testing, and scaling considerations for production environments. Perfect for software engineers, data scientists, and AI practitioners looking to develop production-ready AI applications with a focus on lifecycle management, iterative development, and monitoring systems.

Syllabus

0:00 Introduction and Overview
12:00 Using LMs to Generate Emails and Refining with Feedback
25:00 Transitioning from Pipelines to Human-in-the-Loop Workflows
38:00 Running the App: Setting Up the Environment and Execution
51:00 Extracting Structured Data from PDFs for Email Generation
1:05:00 Evaluating LM Output and Improving Personalization
1:18:00 Testing LM-generated Emails: Metrics and Business Impact
1:32:00 Observability, Tracing, and Error Handling in LM Systems
1:47:00 Building Search and Retrieval Components RAG for LM Apps
2:02:00 Optimizing Prompts and Using Synthetic Data for Testing
2:17:00 Challenges with Prompt Optimization and Fine-tuning
2:32:00 Scaling LMs in Production and Business Considerations
2:45:00 Final Thoughts on AI-driven Email Automation
2:55:00 Next Steps, Q&A, and Closing Discussion