Curated Resources
Every resource hand-picked, explained, and contextualized. Quality over quantity, always.
Machine Learning Specialization
Stanford University / Coursera • Andrew Ng3 months
The definitive introduction to machine learning by the co-founder of Google Brain. This updated 2022 course covers supervised learning, unsupervised learning, and best practices used in Silicon Valley.
Our take: Start here if you're new to ML. Andrew Ng's teaching style makes complex concepts accessible. Over 5 million students have taken this course, making it the gold standard for ML education.
Machine Learning Crash Course
Google Developers • Google AI Team15 hours
Google's fast-paced, practical introduction to machine learning with TensorFlow APIs. Features interactive visualizations, video lectures, and hands-on coding exercises.
Our take: Perfect for developers who want hands-on experience quickly. This is what Google uses internally to train new engineers on ML basics. Completely free with no signup required.
Practical Deep Learning for Coders
fast.ai • Jeremy Howard & Rachel Thomas7 weeks
A top-down approach to deep learning that gets you building state-of-the-art models immediately. Uses PyTorch and the fastai library.
Our take: Unlike traditional courses, fast.ai teaches you to train models first, then understand the theory. Jeremy Howard was president of Kaggle and knows what works in practice.
Neural Networks Series
YouTube • Grant Sanderson (3Blue1Brown)1 hour
Beautiful visual explanations of how neural networks learn, including gradient descent and backpropagation. Uses stunning animations to build intuition.
Our take: The best visual introduction to neural networks ever created. Watch this before diving into code. It builds incredible intuition for how neural networks actually work.
Neural Networks: Zero to Hero
YouTube • Andrej Karpathy10+ hours
Build neural networks from scratch in Python. Former Tesla AI Director and OpenAI researcher teaches you to implement GPT-style models step by step.
Our take: Andrej Karpathy is one of the most respected practitioners in AI. This series shows you how to build everything from micrograd to a GPT. Incredibly educational.
MIT 6.S191: Introduction to Deep Learning
MIT • Alexander Amini & Ava Soleimany10 weeks
MIT's official introductory course on deep learning. Covers foundational concepts, CNNs, RNNs, transformers, and cutting-edge applications.
Our take: A modern, well-produced course that's updated annually. The labs are particularly well-designed and you'll build real deep learning applications.
Attention Is All You Need
arXiv / NeurIPS 2017 • Vaswani et al. (Google Brain)3-4 hours
The foundational paper introducing the Transformer architecture that powers GPT, BERT, and virtually all modern LLMs.
Our take: This 2017 paper revolutionized AI. Understanding the attention mechanism is essential for understanding modern language models. Read it multiple times.
Prompt Engineering Guide
OpenAI • OpenAI Team2 hours
Official guide on prompt engineering and getting the best results from GPT models. Covers strategies, tactics, and best practices.
Our take: Essential reading for anyone working with LLMs. These are the techniques that professionals use to get reliable, high-quality outputs from language models.
Language Models are Few-Shot Learners (GPT-3)
arXiv / NeurIPS 2020 • Brown et al. (OpenAI)4-5 hours
The GPT-3 paper demonstrating that large language models can perform tasks with minimal examples. Introduced the era of few-shot learning.
Our take: This paper kicked off the LLM revolution. Understanding scaling laws and in-context learning is essential for understanding modern AI capabilities.
The Illustrated Transformer
Jay Alammar's Blog • Jay Alammar1 hour
Visual, step-by-step explanation of the Transformer architecture. The clearest explanation of attention mechanisms available online.
Our take: Before reading the original paper, read this. Jay Alammar's visualizations make the Transformer architecture genuinely understandable.
Boston Dynamics: Inside the Lab
YouTube • Boston DynamicsOngoing
Official channel showcasing Atlas, Spot, and the engineering behind the world's most advanced robots. Includes technical talks and demonstrations.
Our take: See what's possible in humanoid robotics. Boston Dynamics represents the cutting edge of physical AI, and their Atlas robot is genuinely remarkable.
Underactuated Robotics
MIT OpenCourseWare • Russ TedrakeFull semester
MIT's course on dynamics, control, and motion planning for underactuated robots. Essential for understanding how robots achieve dynamic, natural movement.
Our take: Russ Tedrake is a leading robotics researcher. This course explains why making robots move gracefully is so challenging and how to solve it.
Tesla Optimus: Humanoid Robot Development
Tesla AI • Tesla Robotics Team30 minutes
Tesla's approach to building general-purpose humanoid robots. Leveraging automotive manufacturing expertise for mass production of humanoids.
Our take: Tesla is betting big on humanoid robots for manufacturing. Their approach to scaling production could transform the industry.
Figure 01: General Purpose Humanoid
Figure AI • Figure AI Team20 minutes
Figure AI's approach to building commercially viable humanoid robots. Backed by major tech investors and partnering with OpenAI for intelligence.
Our take: Figure represents the new wave of humanoid robotics startups. Their partnership with OpenAI shows the convergence of LLMs and robotics.
ROS 2 Documentation & Tutorials
ROS.org • Open RoboticsSelf-paced
Official tutorials for ROS 2 (Robot Operating System), the standard middleware for robot software development used by most research labs and companies.
Our take: ROS is the lingua franca of robotics. Most research labs and companies use it. Learning ROS is essential for any serious robotics work.
Control Bootcamp
YouTube • Steve Brunton, PhD12 hours
Complete introduction to control theory with MATLAB examples. Covers state-space, stability, controllability, and modern control techniques.
Our take: Steve Brunton is an exceptional educator. Understanding control theory is essential for making robots move smoothly, safely, and efficiently.
CS231n: Deep Learning for Computer Vision
Stanford University • Fei-Fei Li, Andrej Karpathy, Justin JohnsonFull quarter
Stanford's legendary computer vision course covering CNNs, object detection, segmentation, and visual recognition. The course that trained a generation of CV engineers.
Our take: Andrej Karpathy's original lectures are exceptional. This course is the foundation for understanding how robots 'see' and interpret the world.
SLAM Course
YouTube • Cyrill Stachniss25+ hours
Comprehensive course on Simultaneous Localization and Mapping (SLAM). Covers graph-based SLAM, particle filters, and visual SLAM techniques.
Our take: SLAM enables robots to navigate unknown environments while building a map. Critical for autonomous humanoids and mobile robots.
Reinforcement Learning for Robotics
Spinning Up (OpenAI) • Josh Achiam (OpenAI)Self-paced
OpenAI's educational resource for deep reinforcement learning. Excellent for understanding how robots can learn complex behaviors through trial and error.
Our take: RL is increasingly important for robot control. Spinning Up is the best starting point: well-organized, well-explained, with working code.
The Future of Jobs Report 2023
World Economic Forum • WEF Research Team2-3 hours
Comprehensive analysis of how technology is reshaping labor markets. Based on surveys of 800+ companies representing 11 million workers across 27 industries.
Our take: The most authoritative annual report on workforce transformation. Essential reading to understand which skills are growing, declining, and emerging globally.
The State of AI in 2024
McKinsey Global Institute • McKinsey Research1-2 hours
Annual survey of AI adoption across industries. Reveals how organizations are actually using AI and what skills they're hiring for.
Our take: McKinsey surveys thousands of organizations annually. This shows the gap between AI hype and actual enterprise adoption, helping you focus on skills that matter.
Google AI Essentials
Google / Coursera • Google Career Certificates Team10 hours
Beginner-friendly course designed for non-technical professionals. Learn to use AI tools effectively in everyday work without coding.
Our take: Perfect starting point if you're not technical but need to understand AI. Google designed this for their own non-technical employees first.
AI Skills for the Modern Workplace
LinkedIn Learning • LinkedIn Economic Graph Team1 hour
Data-driven analysis of AI skill trends across 900 million professionals. Shows which AI skills are being added to profiles and which jobs are adding AI requirements.
Our take: LinkedIn has unique data on actual skill trends. This report shows what professionals are actually learning, not just what experts say they should learn.
Elements of AI
University of Helsinki • Reaktor & University of Helsinki6 weeks
Free online course designed to demystify AI for everyone. No programming required. Over 1 million students from 170 countries have completed it.
Our take: The most accessible AI introduction available. If technical courses feel intimidating, start here. It builds genuine understanding without overwhelming.
Prompt Engineering for Everyone
DeepLearning.AI • Andrew Ng & Isa Fulford1 hour
Practical course on getting better results from AI language models. Learn principles and tactics that work across different AI tools.
Our take: Prompt engineering is becoming a universal skill. This free course from Andrew Ng teaches the techniques that actually make AI tools more useful.
Natural Language Processing with Deep Learning (CS224N)
Stanford University • Prof. Christopher Manning10 weeks
Stanford's flagship NLP course covering word vectors, transformers, pretraining, and large language models. Includes assignments in PyTorch.
Our take: The gold standard for NLP education. Prof. Manning is a co-inventor of key NLP techniques. Free lecture videos, slides, and assignments available online.
OpenAI Academy
OpenAI • OpenAI TeamSelf-paced
Official learning platform from OpenAI covering prompting techniques, API usage, fine-tuning, and building with GPT models. From beginner to advanced.
Our take: Learn directly from the creators of ChatGPT and GPT-4. The most authoritative resource for understanding and building with OpenAI's technology stack.
CS229: Machine Learning
Stanford University • Andrew NgFull quarter
Stanford's famous graduate-level machine learning course. More mathematically rigorous than the Coursera version, covering the theoretical foundations of ML algorithms.
Our take: The full Stanford experience. If you want to truly understand the math behind ML, this is the course. Lecture videos and notes freely available.
StatQuest Machine Learning Playlist
YouTube • Josh Starmer, PhDSelf-paced
Clear, visual explanations of machine learning concepts with memorable songs and animations. Covers everything from linear regression to XGBoost.
Our take: Josh Starmer has a gift for making statistics intuitive. When you're confused about a concept, StatQuest probably has a video that will make it click.
Deep Learning Textbook
MIT Press • Ian Goodfellow, Yoshua Bengio, Aaron CourvilleSelf-paced
The comprehensive textbook on deep learning, freely available online. Written by pioneers including the inventor of GANs.
Our take: The 'bible' of deep learning. Dense but thorough, best used as a reference alongside practical courses. Essential for anyone serious about understanding deep learning theory.
Deep Learning with PyTorch
PyTorch.org • PyTorch Team60 minutes
Official PyTorch tutorial covering tensors, autograd, neural networks, and training classifiers. The fastest way to get productive with PyTorch.
Our take: PyTorch is now the dominant framework in AI research. This tutorial gets you writing real neural networks in under an hour. Bookmark this.
NLP Course
Hugging Face • Hugging Face TeamSelf-paced
Comprehensive course on using Transformers for NLP tasks. Learn to fine-tune models, build applications, and use the Hugging Face ecosystem.
Our take: Hugging Face is the GitHub of ML models. This course teaches you to use their ecosystem effectively, which is essential knowledge for working with LLMs in production.
BERT: Pre-training of Deep Bidirectional Transformers
arXiv / NAACL 2019 • Devlin et al. (Google AI)2-3 hours
The paper that introduced BERT, revolutionizing NLP by showing how bidirectional pre-training dramatically improves language understanding.
Our take: BERT changed how we approach NLP. Understanding masked language modeling and fine-tuning is crucial for modern NLP work.
LLaMA: Open and Efficient Foundation Language Models
arXiv • Touvron et al. (Meta AI)2-3 hours
Meta's open-source LLM that matches GPT-3 performance with fewer parameters. The paper that democratized large language model research.
Our take: LLaMA opened up LLM research to the broader community. Understanding efficient training and model architecture choices is increasingly important.
IEEE Humanoid Robots Conference Papers
IEEE Xplore • Various ResearchersVariable
Archive of research papers from IEEE-RAS International Conference on Humanoid Robots. Cutting-edge research from top robotics labs worldwide.
Our take: The IEEE Humanoids conference is where researchers present breakthrough work. Great for diving deep into specific technical challenges.
Model Predictive Control Tutorial
YouTube • MATLAB5 hours
Understanding Model Predictive Control (MPC), the control strategy used in many advanced robotic systems including Boston Dynamics' robots.
Our take: MPC is increasingly used in cutting-edge robotics. This series explains why it's so powerful for handling constraints and predictions.
Legged Locomotion: Models, Concepts, and Robots
MIT Press • Various AuthorsSelf-paced
Comprehensive coverage of legged robot locomotion, from simple models to complex humanoid walking. A foundational text in the field.
Our take: Walking is deceptively hard for robots. This book explains the physics and control strategies that make it possible.
OpenCV Python Tutorial
OpenCV.org • OpenCV TeamSelf-paced
Official OpenCV tutorials for Python. Learn image processing, feature detection, object tracking, and camera calibration.
Our take: OpenCV is the most widely used computer vision library. These tutorials cover everything you need for practical robot vision applications.
Monocular Depth Estimation: A Survey
arXiv • Various Researchers3-4 hours
Comprehensive survey of monocular depth estimation methods. Covers classical approaches through modern deep learning techniques.
Our take: Depth perception from a single camera is increasingly important for robotics. This survey covers the state of the art.
Sensor Fusion and Kalman Filtering
YouTube • Michel van Biezen8 hours
Clear explanations of Kalman filtering for sensor fusion. Essential for combining data from multiple sensors (IMU, cameras, LiDAR) in robotics.
Our take: Real robots need to combine information from multiple sensors. Kalman filtering is the fundamental technique for doing this reliably.
The Elements of Statistical Learning
Stanford University • Hastie, Tibshirani, FriedmanSelf-paced
The definitive text on statistical learning theory. Freely available PDF from the authors. Covers everything from linear regression to neural networks.
Our take: A classic that every ML practitioner should have on their shelf (or browser). More mathematical than practical courses, but essential for deep understanding.
Understanding LSTM Networks
Colah's Blog • Christopher Olah30 minutes
The clearest explanation of LSTM networks and recurrent neural networks. Beautiful diagrams that make complex architectures intuitive.
Our take: Chris Olah is famous for his ability to explain complex neural network concepts. This article has helped millions understand RNNs and LSTMs.
Scaling Laws for Neural Language Models
arXiv • Kaplan et al. (OpenAI)2-3 hours
The paper that revealed predictable scaling laws for language models. Shows how performance improves with model size, data, and compute.
Our take: Understanding scaling laws is crucial for understanding why bigger models keep getting better, and the economics of AI development.
Agility Robotics: Digit Humanoid
Agility Robotics • Agility Robotics Team30 minutes
Agility Robotics' Digit is one of the first humanoid robots being deployed commercially in warehouses. Learn about their design philosophy and deployment strategy.
Our take: Digit represents the practical, near-term future of humanoid robots. Understanding their design choices reveals what's actually deployable today.
NeRF: Neural Radiance Fields
arXiv / ECCV 2020 • Mildenhall et al. (UC Berkeley)3-4 hours
Revolutionary technique for synthesizing novel views of complex scenes. NeRFs are transforming how robots understand and represent 3D environments.
Our take: NeRFs represent a paradigm shift in 3D scene understanding. Understanding this technique is increasingly important for robot perception.
How to Become an AI-First Company
Harvard Business Review • Thomas H. Davenport & Nitin Mittal30 minutes
Strategic framework for integrating AI into organizations. Explains the skills and mindsets needed at every level of the organization.
Our take: HBR provides the business context that technical resources miss. Understanding how leaders think about AI helps you position your skills strategically.
Artificial Intelligence: Implications for Business Strategy
MIT Sloan • MIT Sloan Executive Education6 weeks
Executive-level understanding of AI capabilities, limitations, and strategic implications. Designed for leaders making AI investment decisions.
Our take: If you're in a leadership role or aspire to one, understanding AI strategy is essential. MIT Sloan provides the frameworks executives actually use.
Jobs Lost, Jobs Gained: Workforce Transitions
McKinsey Global Institute • James Manyika et al.2 hours
Landmark research on how automation will affect jobs through 2030. Analyzes 800 occupations across 46 countries to identify which tasks are automatable.
Our take: The most rigorous analysis of job automation risk. Helps you understand which parts of your role are vulnerable and which are not.
Microsoft AI Skills Initiative
Microsoft • Microsoft Learn TeamSelf-paced
Free learning paths covering AI fundamentals, generative AI, and responsible AI practices. Includes hands-on experience with Microsoft AI tools.
Our take: Microsoft's AI tools are widely used in enterprises. Understanding Copilot, Azure AI, and their ecosystem is practical knowledge for many workplaces.
OECD Skills Outlook: Skills for the Digital Transition
OECD • OECD Research Team2-3 hours
International analysis of how digitalization is changing skill demands across OECD countries. Evidence-based policy recommendations for lifelong learning.
Our take: OECD provides the international perspective. Understanding global skill trends helps you see beyond local job market fluctuations.
AI Career Guide: The AI Ladder
IBM • IBM Skills Team30 minutes
Framework for understanding AI maturity in organizations and careers. Helps identify where you are and what skills to develop next.
Our take: IBM's AI Ladder framework is widely used in enterprise. Understanding it helps you speak the language that organizations use when discussing AI adoption.
Generative AI Learning Path
Google Cloud • Google Cloud Training8 hours
Comprehensive learning path covering generative AI fundamentals, large language models, responsible AI, and practical applications with Google's AI tools.
Our take: Google's structured approach to GenAI education. Includes hands-on labs and quizzes. Earn skill badges to validate your knowledge.
Hugging Face NLP Course
Hugging Face • Hugging Face Team20 hours
Free, practical course on NLP using the Transformers library. Covers tokenization, fine-tuning pretrained models, and building NLP applications.
Our take: Hugging Face is the de facto standard for sharing and using pretrained models. This course teaches you to use their ecosystem which powers most modern NLP work.
DeepMind x UCL Deep Learning Lecture Series
DeepMind / University College London • DeepMind Research Scientists12 lectures
Lecture series from leading DeepMind researchers covering optimization, CNNs, attention, generative models, graph neural networks, and unsupervised learning.
Our take: Taught by the scientists behind AlphaGo and AlphaFold. Offers a rare window into how top AI researchers think about deep learning problems.
Convolutional Neural Networks for Visual Recognition (CS231N)
Stanford University • Fei-Fei Li, Andrej Karpathy10 weeks
Stanford's legendary computer vision course. Covers image classification, object detection, neural network architectures, and visual understanding with deep learning.
Our take: Created by Fei-Fei Li (co-creator of ImageNet) and Andrej Karpathy (former Tesla AI director). The assignments are challenging but transform your understanding of vision AI.
Reinforcement Learning Lecture Series
DeepMind / University College London • Hado van Hasselt, Diana Borsa13 lectures
Comprehensive lecture series on reinforcement learning from DeepMind researchers. Covers MDPs, policy gradient methods, model-based RL, and multi-agent systems.
Our take: The definitive RL course from the team that built AlphaGo, AlphaStar, and AlphaFold. Essential viewing for anyone serious about RL.
* Affiliate Disclosure: Some links may earn us a commission at no extra cost to you. We only recommend resources we genuinely believe in.
Sponsored
Recommended Books
Hand-picked books from industry experts. Only the best resources that we genuinely recommend.
Affiliate Disclosure: Some links may earn us a commission at no extra cost to you.
Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow
The most practical guide to ML. Covers everything from linear regression to deep neural networks with hands-on code examples.
Industry standard for practical ML. Used by Google engineers and top bootcamps.
Topics covered:
Deep Learning
The definitive textbook on deep learning. Comprehensive coverage of mathematical foundations and modern techniques.
Written by pioneers of the field. The "bible" of deep learning used in top universities.
Topics covered:
Pattern Recognition and Machine Learning
Rigorous mathematical treatment of ML algorithms. Essential for understanding the theory behind modern methods.
Gold standard for ML theory. Required reading at Cambridge, Stanford, and MIT.
Topics covered:
Artificial Intelligence: A Modern Approach
The most comprehensive AI textbook covering search, planning, reasoning, learning, and perception.
Used in 1,500+ universities worldwide. Written by Google Director of Research.
Topics covered:
Speech and Language Processing
Comprehensive introduction to NLP and computational linguistics. Covers classical and neural approaches.
Stanford standard NLP textbook. Authors are leading researchers in the field.
Topics covered:
Natural Language Processing with Transformers
Practical guide to using Hugging Face Transformers for NLP tasks. From text classification to question answering.
Written by Hugging Face engineers. The go-to guide for modern NLP with transformers.
Topics covered:
Machine Learning Design Patterns
Solutions to common challenges in ML systems. Covers data representation, problem framing, and production deployment.
Written by Google Cloud AI engineers. Bridges the gap between ML theory and production.
Topics covered:
Building Machine Learning Pipelines
Automate model life cycles with TensorFlow Extended (TFX). From data validation to serving.
Essential for MLOps and deploying ML models at scale. Used by production ML teams.
Topics covered:
Probabilistic Robotics
The standard textbook for robot perception and navigation. Covers localization, mapping, and SLAM.
Written by the founder of Google X and Waymo. Essential for autonomous systems.
Topics covered:
Modern Robotics: Mechanics, Planning, and Control
Modern approach to robot kinematics, dynamics, and control with accompanying video lectures.
Accompanies popular Coursera specialization. Clear explanations with practical focus.
Topics covered:
Robotics, Vision and Control
Practical introduction to robotics with MATLAB examples. Covers vision, arms, and mobile robots.
Excellent for hands-on learners. Comes with MATLAB Robotics Toolbox.
Topics covered:
Python Machine Learning
Practical introduction to ML with Python. Great for beginners transitioning from programming to ML.
Best selling ML book for Python developers. Excellent code examples.
Topics covered:
Fluent Python
Master Python's most powerful features. Essential for writing clean, efficient ML code.
Best Python book for intermediate developers. Makes you a better Python programmer.
Topics covered:
Reinforcement Learning: An Introduction
The definitive introduction to reinforcement learning. Covers theory and algorithms comprehensively.
Written by the pioneers of RL. Free PDF available but physical copy recommended for study.
Topics covered:
Deep Learning for Computer Vision
Practical guide to CNN architectures for image classification, detection, and segmentation.
Hands-on approach to computer vision with deep learning. Great code examples.
Topics covered:
Programming Computer Vision with Python
Hands-on introduction to computer vision using Python. From basics to 3D reconstruction.
Classic introduction to CV with Python. Great for building foundations.
Topics covered:
Mathematics for Machine Learning
Essential mathematics for understanding ML algorithms. Linear algebra, calculus, and probability.
Free PDF available. Perfect bridge from math basics to ML. Written by Imperial College professors.
Topics covered:
Generative Deep Learning
Comprehensive guide to generative models including VAEs, GANs, Transformers, and Diffusion models.
Most up-to-date book on generative AI. Covers latest architectures including GPT and Stable Diffusion.
Topics covered:
Building LLM Apps
Practical guide to building applications with large language models using LangChain and OpenAI APIs.
Perfect for developers wanting to build with GPT-4, Claude, and other LLMs.
Topics covered:
Build a Large Language Model (From Scratch)
Step-by-step guide to building a GPT-like LLM from the ground up. Covers tokenization, attention mechanisms, pretraining, and fine-tuning with clear code.
The only book that walks you through building an LLM from absolute scratch. Transforms your understanding of how ChatGPT-style models actually work.
Topics covered:
Designing Machine Learning Systems
An iterative process for designing production ML systems. Covers data engineering, model development, deployment, monitoring, and responsible AI.
Written by a Stanford lecturer and industry veteran. The bridge between academic ML and real-world production systems that most courses skip.
Topics covered:
AI for Robotics: Toward Embodied Intelligence
Approaches robotics from a deep learning perspective. Covers embodied AI, perception, manipulation, navigation, and human-robot interaction.
One of the first books to bridge modern deep learning with practical robotics. Essential reading for anyone at the intersection of AI and physical systems.
Topics covered: