Learning Roadmap: GenAI engineering

GenAI engineering sits at the intersection of large language models, prompt design, retrieval-augmented generation (RAG), evaluation, deployment, and increasingly AI agents. The best video-first courses for this topic are the ones that combine conceptual grounding with practical engineering workflows: how LLMs work, how to build useful applications around them, how to evaluate outputs, and how to move from toy demos to production-ready systems. The strongest free options currently come from reputable providers that actively build in this space, including DeepLearning.AI, Hugging Face, Microsoft, Google, fast.ai, Coursera, and Kaggle. ([coursera.org](https://www.coursera.org/learn/generative-ai-with-llms?irgwc=1)) These recommendations were chosen because they are genuinely free or explicitly enrollable for free, are currently available, and cover the full GenAI engineering arc from beginner-friendly foundations to implementation details like fine-tuning, evaluation, agent frameworks, deployment, and practical tooling. Together, they form a strong self-study curriculum: start with broad intuition, then learn LLM mechanics, then application patterns such as prompting and RAG, and finally move into agentic systems and production-oriented engineering practices. ([deeplearning.ai](https://www.deeplearning.ai/alpha/courses/generative-ai-for-everyone/))

Prerequisites: For beginner courses, no formal prerequisites are required, though basic comfort with computers and web apps helps. For the more technical courses, you should ideally know basic Python and have some familiarity with machine learning concepts like training, validation, and supervised learning. If you do not yet have that background, start with Generative AI for Everyone and Google’s Machine Learning Crash Course before the intermediate offerings. ([deeplearning.ai](https://www.deeplearning.ai/alpha/courses/generative-ai-for-everyone/))

Estimated time: Approximately 95-120 hours total

Start with Generative AI for Everyone to build intuition and vocabulary, then take Machine Learning Crash Course if your ML foundation is weak. Next, move into Getting Started with Generative AI for application patterns like prompting, vector databases, RAG, and agents, followed by Generative AI for Beginners from Microsoft for a more developer-oriented implementation path. After that, take Generative AI with Large Language Models to understand the full LLM lifecycle, including fine-tuning and evaluation. Then specialize with the Hugging Face AI Agents Course to learn modern agentic architectures and tool use. Use the Kaggle 5-Day Gen AI Intensive Course as a quick practical refresher along the way, and complete Practical Deep Learning for Coders if you want deeper long-term engineering strength in modeling, experimentation, and deployment.