Advanced Generative AI Applications

This unit explores advanced generative model architectures and practical techniques for adapting generative AI to real-world creative and analytical tasks. Students will learn feature engineering, model customization (VAEs, flows, GANs, hybrid models), prompt and iteration strategies, and evaluation practices—while addressing common pitfalls, ethical considerations, and content repurposing for media and time-series forecasting.

Learning Objectives

• Analyze and compare advanced generative architectures (VAEs, normalizing/continuous flows, autoregressive flows, WGANs, and hybrid models) and select appropriate approaches for given creative or time-series problems
• Apply feature-engineering and model-customization techniques to adapt LLMs and generative models for domain-specific tasks such as content generation, logo/slogan design, SEO optimization, and next-day return forecasting
• Demonstrate effective prompt sequencing, iteration strategies, and collaborative workflows with AI to produce coherent, concise, and ethically responsible multi-author content across text, social media, and creative assets
• Evaluate experimental results and model outputs to identify common pitfalls, separate fact from fiction, implement improvements, and document recommendations for deployment and continued refinement