Advanced: Generative AI for Developers
Google Skills
Last updated September 30, 2025
A technical, under-the-hood generative-AI path — about 21 hours — built for app developers, machine-learning engineers, and data scientists. It works through image generation with diffusion models, attention mechanisms, encoder-decoder and Transformer architectures, image-captioning models, Vertex AI Studio, vector search and embeddings, multimodal RAG with Gemini, responsible AI, and MLOps for generative AI. Concept material is free; the hands-on labs run on Google Skills and need a subscription or credits. The recommended prerequisite is the Introduction to Generative AI path, and the work is hands-on coding.
What you'll learn
- Image generation with diffusion models and building image-captioning models
- Attention mechanisms and encoder-decoder architectures
- Transformer and BERT models
- Vector search and embeddings
- Inspecting rich documents with Gemini multimodality and multimodal RAG
- Responsible AI for developers and MLOps for generative AI
Frequently asked questions about Advanced
Who is Advanced for?
For app developers, machine-learning engineers, and data scientists who want a technical, hands-on understanding of how generative-AI models work and how to build with them on Vertex AI.
What are the prerequisites for Advanced?
A developer background; the Introduction to Generative AI path is the recommended prerequisite.
Do you need to code for Advanced?
Yes — Advanced involves hands-on coding.
Why we suggest this course
This is the technical complement to the conceptual gen-AI material: it opens up how the models actually work — diffusion, attention, Transformers, embeddings, multimodal RAG — and has you build with them in Vertex AI rather than just describe them. One thing to know: it is genuinely advanced and assumes a developer background, with the Introduction to Generative AI path recommended first to make the most of it.