“What are the best platforms for learning about AI?”

Here are strong, up-to-date platforms to learn AI, grouped by what you might need:

Beginner-friendly, structured paths

• DeepLearning.AI on Coursera or learn.deeplearning.ai: Andrew Ng’s Machine Learning and Deep Learning Specializations; newer agentic/LLM courses are on DeepLearning.AI’s own platform. (coursera.org)
• Google’s Machine Learning Crash Course: free, hands-on intro with videos and exercises. (developers.google.com)
• Kaggle Learn: short, practical micro-courses (Python, ML, model eval) plus competitions to practice. (kaggle.com)

Hands-on modern NLP/LLMs

• Hugging Face Course: practical training on transformers, datasets, tokenizers, and the Hub. (huggingface.co)
• OpenAI Academy: guided courses and badges focused on building with OpenAI tools (useful if you plan to ship with OpenAI). (openai.com)

University-level depth (free/low-cost)

• Stanford CS229 (Machine Learning): lectures and materials online; rigorous foundations. (youtube.com)
• MIT 6.S191 (Intro to Deep Learning) via MIT OpenCourseWare. (ocw.mit.edu)
• edX AI/ML programs from universities and industry (audit many courses free). (edx.org)

Professional/enterprise skilling

• NVIDIA Deep Learning Institute: instructor-led and self-paced courses with certificates (great if you use NVIDIA stacks). (nvidia.com)
• Microsoft Learn: learning paths for Azure AI, agents, and applied AI scenarios. (learn.microsoft.com)

Math foundations (to make the ML “click”)

• Khan Academy (algebra, linear algebra refreshers). (khanacademy.org)

Quick picks by goal

• “I want a single guided start”: DeepLearning.AI Machine Learning Specialization, then Kaggle Learn exercises. (coursera.org)
• “I want to build with LLMs now”: Hugging Face Course + OpenAI Academy (if targeting OpenAI stack). (huggingface.co)
• “I want theory depth for grad-level understanding”: Stanford CS229 + MIT 6.S191. (cs229.stanford.edu)

If you share your background (coding/math experience) and goals (research, apps, career switch), I can map this into a sequenced plan with time estimates and a few project ideas.