StanfordAutumn 20267 lectures · 7h 41m
CS193T: Thinking with AI
Cynthia Bailey
A hands-on course with no prerequisites on working with AI tools, covering prompting, agents, evaluating models and knowing when to push back.
Resources
The material I learn from on LLMs, agents, efficient AI and running systems. Nearly all of it is free, and each entry says in a line what it covers.
A few of these in the order I'd take them, with a line on why each comes where it does.
3 steps · 46h of lectures
How a Transformer becomes an LLM, how to make one cheap enough to run, and the hardware and energy it runs on.
4 steps · 21h of lectures
From using AI tools well to engineering agents and building software with them.
11 of 11 resources
StanfordAutumn 20267 lectures · 7h 41m
Cynthia Bailey
A hands-on course with no prerequisites on working with AI tools, covering prompting, agents, evaluating models and knowing when to push back.
StanfordAutumn 202610 lectures · 13h
Diyi Yang, Michael Ryan, John Yang
Building agentic systems, from simple LLM pipelines to compound systems and autonomous agents, with RAG, tool use, DSPy, evaluation and a quarter-long project.
Lecture notes with Python code2026
Stan Birchfield
Sixty-six lessons with working Python code, from classical image processing and 3D vision to deep learning and foundation models. Also a book on arXiv.
StanfordAutumn 20267 lectures · 12h
Afshine Amidi, Shervine Amidi
From the evolution of NLP to the core Transformer components, how they become LLMs, and the techniques that make LLMs perform better.
StanfordAutumn 2026
Mihail Eric
AI-native software development, covering MCP, agent skills, spec-driven development and building a software factory.
Harvard
Pranav Rajpurkar
The working toolkit of AI research, from PyTorch, Lightning and Hugging Face to reading papers, finding ideas and presenting results.
MIT Han LabFall 202619 lectures · 21h
Song Han
Making models cheap to run, from pruning, quantization and distillation to distributed training, serving and acceleration for LLMs and diffusion models.
StanfordSpring 202613 lectures · 13h
Anjney Midha, Michael Abbott
The stack behind frontier AI, from energy and silicon to models, applications, security and deployment policy, with weekly talks from people building it.
Google2020
Heather Adkins, Betsy Beyer and 4 others
Designing, building and running systems that stay secure and reliable, through incident response, recovery and the culture that keeps it working. Free online.
Google2016
Betsy Beyer, Chris Jones and 2 others
How Google runs production, with SLOs, monitoring, on-call, incident management and postmortems. Free online.
Stanford CS146SVideo
Lee Robinson
A guest lecture in CS146S on how agents that keep running in the background are built.