Artificial Intelligence A-Z 2025: Agentic AI, Gen AI, and RL
-
Welcome to the course!
-
Agentic AI
-
——–Part 0 – Fundamentals Of Reinforcement Learning——–
-
Q-Learning IntuitionDeep Learning Fundamentals: Neural Networks & Activation Functions ExplainedHow Reinforcement Learning Works: A Beginner’s Guide to AI Training MethodsBellman Equation in Reinforcement Learning: A Step-by-Step IntroductionFrom State Values to Optimal Plans: Bellman Equation in AI Decision MakingMarkov Decision Processes in Reinforcement Learning: A Complete GuideRL Tutorial: Optimal Policy vs Fixed Plans in AI Decision MakingLiving Penalty in Reinforcement Learning: Optimize AI Agent Decision MakingQ-Learning in Reinforcement Learning: From V-Values to Q-Values ExplainedTemporal Difference in Q-Learning: A Complete Guide for Reinforcement LearningQuiz 1
-
Q-Learning Implementation
-
Part 1 – Deep Q-Learning
-
Deep Q-Learning IntuitionDeep Learning Fundamentals: Neural Networks & Activation Functions ExplainedDeep Q-Learning vs Traditional Q-Learning: Key Differences ExplainedHow Deep Q-Learning Works: Neural Networks & Reinforcement Learning ExplainedExperience Replay in Deep Q-Learning: How it Works & Why it MattersQ-Learning: Guide to Epsilon-Greedy & Softmax Action Selection Algorithms
-
Deep Q-Learning ImplementationHow to Train an AI Lunar Lander using Deep Q-Learning and PythonGet the Codes hereStep 1 – Deep Q-Learning Environment Setup: From Gmail to Lunar Lander TrainingGoogle Colab Setup: Deep Q-Learning for Lunar Lander TutorialStep 3 – PyTorch DQN Architecture: Building the AI Brain for OpenAI Lunar LanderPyTorch Deep Q-Learning: Implementing Forward Method for Neural NetsStep 5 – Configure LunarLander-v2 Environment Parameters for DQN TrainingDQN Hyperparameters: Learning Rate & Replay Buffer Setup Guide (Step 6)Step 7: Implementing Experience Replay Memory in DQN with PythonStep 8: DQN Push Method – Adding Experiences to Replay Memory BufferStep 9: Coding DQN Memory Sampling – PyTorch Experience Replay TutorialDQN Tutorial: Initialize Q-Networks, Optimizer & Replay Memory BufferStep 11: DQN Step Method – Store & Learn from Experiences in PythonStep 12: DQN Action Selection – State Processing to Policy ImplementationStep 13: Deep Q-Network Training – Implementing Learn Method for RLStep 14 – Deep Q-Network Implementation: Soft Update Method for Stable TrainingStep 15: Creating Your First AI Agent – Deep Q-Network (DQN) TutorialStep 16 – Epsilon-Greedy Strategy: Initializing AI Training HyperparametersStep 17: Deep Q-Learning Training Loop – Complete Lunar Lander GuideStep 18: DQN Training Visualization – Dynamic Score Tracking ImplementationStep 19: Visualizing Deep Q-Learning – AI Perfects Lunar Lander LandingChatGPT vs Custom DQN: Comparing Deep RL Implementations
-
——–Part 2 – Deep Convolutional Q-Learning——–
-
Deep Convolutional Q-Learning Intuition
-
Deep Convolutional Q-Learning ImplementationDeep Q-Learning: Build AI Pac-Man with Convolutional Neural NetworksGet the Codes hereDeep Q-Learning: Setting Up Pac-Man in OpenAI Gym – Step 1Step 2 – Implementing DCQN Architecture in Python: Setup & Neural Network DesignStep 3: Building CNNs – Creating AI’s Visual Processing SystemCNN Architecture: Adding Fully Connected Layers After Convolutional LayersStep 5: Deep Q-Learning – Building AI’s Visual Processing SystemStep 6: Configuring Miss Pacman for Deep Q-Learning TrainingStep 7: Deep Q-Learning Hyperparameters – Learning Rate & Batch Size SetupImage Preprocessing for Deep Q-Learning: PIL & Torchvision ImplementationStep 9: Deep Q-Learning to DCQN – Experience Replay & Memory UpdatesStep 10: Implementing DCQN Agent – Deep Q-Learning Adaptation & MethodsStep 11: Optimizing DQN Training on V100 GPU – Setup to Solved EnvironmentStep 12: Visualizing Deep Q-Learning – Watch AI Play Pac-Man Like a HumanStep 13: Deep Q-Learning – Optimizing Neural Networks with GPT-4
-
——–Part 3 – A3C——–
-
A3C IntuitionDeep Learning Fundamentals: Neural Networks & Activation Functions ExplainedA3C Algorithm Tutorial: Understanding Asynchronous Advantage Actor-Critic in AIActor-Critic Algorithm: From Deep Q-Learning to A3C ImplementationAsynchronous Learning in A3C: Shared Critics and Neural Networks ExplainedHow Does Advantage Work in Actor-Critic Methods? A3C Algorithm ExplainedLSTM in A3C Algorithm: How Memory Enhances Reinforcement Learning Performance
-
A3C ImplementationA3C Reinforcement Learning Tutorial: Building AI for Kung Fu Master EnvironmentGet the Codes hereA3C Deep RL: Setting Up Kung Fu Master Environment in AtariA3C Algorithm Implementation: Neural Network & Environment SetupStep 3 – A3C Deep Reinforcement Learning: Network Class & Architecture DesignA3C Algorithm: Building Action & State Value Outputs – Step 4A3C Algorithm: PreprocessAtari Class & Hyperparameter Tuning SetupStep 6: A3C Agent Class Init Method for Deep Reinforcement LearningA3C Agent: Converting States to Actions with PyTorch Neural NetworksStep 8: Coding A3C Step Method in PyTorch – Complete Implementation GuideStep 9 – How to Initialize an A3C Agent in Python: Creating the Agent InstanceImplementing A3C Agent Evaluation in Python | Deep RL Tutorial Step 10Step 11: EnvBatch Class for A3C Multi-Environment Training in PythonStep 12: Multi-Environment A3C – EnvBatch Class Implementation TutorialStep 13: A3C Training – Multi-Environment Batch Setup ImplementationStep 14: A3C Training Loop with Progress Bar for Kung Fu AI ModelChatGPT A3C Model: PyTorch for KungFuMaster AI Optimization (Step 15)
-
——–Part 4 – PPO and SAC——–
-
——–Part 5 – Intro to Large Language Models (LLMs)——–
-
LLMs IntuitionIntroduction to Large Language Models (LLMs): Transformers ExplainedBuilding Large Language Models: Essential Ingredients for LLM DevelopmentHow Were Large Language Models Invented? Origins of Transformer AIUnderstanding Next Word Prediction: How LLMs Process Text One Word at a TimeHow Do Large Language Models Work? A Deep Dive into LLM ArchitectureWhat Are LLM Parameters? Understanding Large Language Model Size ExplainedContext Windows Explained: How LLMs Remember Conversation HistoryFine-Tuning Large Language Models: Real-World Applications and Use Cases
-
LLMs ImplementationFine-Tuning LLMs for Medical Chatbots: A Practical Guide with Hugging FaceGet the Codes hereInstalling & Importing Key Libraries for Fine-Tuning Llama 2 ModelsLoading LLaMA 2 Model: Hugging Face Transformers & 4-bit PrecisionLoading & Configuring HuggingFace Tokenizer for LLaMA 2 ImplementationStep 4 – Setting Training Arguments for LLM Fine-Tuning in Transformers LibraryStep 5 – Implementing SFTTrainer: Memory-Efficient LLM Training with PEFT & LoRAStep 6: LoRA & Quantization for Efficient LLM Training in Medical TermsStep 7: Chatting with Your Fine-Tuned Medical LLM via Text GenerationBuild a RAG-powered Generative AI application with Knowledge Bases, from scratch
-
THANK YOU
-
Annex 1: Artificial Neural NetworksWhat is Deep Learning? A Beginner’s Guide to Artificial Neural NetworksDeep Learning Fundamentals: Neural Networks & Activation Functions ExplainedHow Do Artificial Neurons Work? A Complete Guide to Neural Network BasicsNeural Network Activation Functions: ReLU, Sigmoid, Tanh & Threshold ExplainedHow Do Neural Networks Work? A Step-by-Step Property Valuation ExampleHow Do Neural Networks Learn? Understanding Backpropagation & Cost FunctionsUnderstanding Gradient Descent: Optimize Neural Network Weights EfficientlyStochastic Gradient Descent vs Batch Gradient Descent: What’s the Difference?Backpropagation in Neural Networks: Step-by-Step Training Guide
-
Annex 2: Convolutional Neural NetworksDeep Learning Fundamentals: Neural Networks & Activation Functions ExplainedCNN vs Human Vision: How Convolutional Neural Networks Process ImagesCNN Convolution: Feature Detection & Feature Maps in Deep LearningReLU in CNN: Understanding Non-Linearity for Deep LearningStep 2 – Max Pooling in CNN: How to Reduce Feature Maps & Prevent OverfittingStep 3 – CNN Feature Map Flattening: From Pooling Layer to Neural Network InputStep 4: CNN Classification – How FC Layers Process FeaturesCNN Architecture Explained: Feature Detection to Neural Network ClassificationUnderstanding Softmax and Cross-Entropy Loss Functions in Neural Networks
-
Congratulations!! Don’t forget your Prize :)
Welcome to Artificial Intelligence A-Z!
Learn key AI concepts with intuition lectures to get you quickly up to speed with all things AI and practice them by building 8 different AIs:
Build an AI Agent with a Foundation Model (LLM) for business assistance, all powered by the Cloud.
Build an AI with a Q-Learning model and train it to optimize warehouse flows in a Process Optimization case study.
Build an AI with a Deep Q-Learning model and train it to land on the moon.
Build an AI with a Deep Convolutional Q-Learning model and train it to play the game of Pac-Man.
Build an AI with an A3C (Asynchronous Advantage Actor-Critic) model and train it to fight Kung Fu.
Build an AI with a PPO (Proximal Policy Optimization) model and train it for a Self-Driving Car.
Build an AI with a SAC (Soft Actor-Critic) model and train it for a Self-Driving Car.
Build an AI by fine-tuning a powerful pre-trained LLM (Llama 2 by Meta) with Hugging Face and re-train it to chat with you about medical terms. Simply put, we build here an AI Doctor Chatbot.
But that’s not all… Once you complete the course, you will get 3 extra AIs: DDPG, Full World Model, and Evolution Strategies & Genetic Algorithms. We build these AIs with ChatGPT for a Self-Driving Car and a Humanoid application. For each of these extra AIs you will get a long video lecture explaining the implementation, a mini PDF, and the Python code.
Besides, you will get a free 3-hour extra course on Generative AI and LLMs with Cloud Computing as a Prize for completing the course.
And last but not least, here is what you will get with this course:
1. Complete beginner to expert AI skills – Learn to code self-improving AI for a range of purposes. In fact, we code together with you. Every tutorial starts with a blank page and we write up the code from scratch. This way you can follow along and understand exactly how the code comes together and what each line means.
2. Hassle-Free Coding and Code templates – We will build all our AIs in Google Colab, which means that we will have absolutely NO hassle installing libraries or packages because everything is already pre-installed in Google Colab notebooks. Plus, you’ll get downloadable Python code templates (in .py and .ipynb) for every AI you build in the course. This makes building truly unique AI as simple as changing a few lines of code. If you unleash your imagination, the potential is unlimited.
3. Intuition Tutorials – Where most courses simply bombard you with dense theory and set you on your way, we believe in developing a deep understanding for not only what you’re doing, but why you’re doing it. That’s why we don’t throw complex mathematics at you, but focus on building up your intuition in AI for much better results down the line.
4. Real-world solutions – You’ll achieve your goal in not only one AI model but in 5. Each module is comprised of varying structures and difficulties, meaning you’ll be skilled enough to build AI adaptable to any environment in real life, rather than just passing a glorified memory “test and forget” like most other courses. Practice truly does make perfect.
5. In-course support – We’re fully committed to making this the most accessible and results-driven AI course on the planet. This requires us to be there when you need our help. That’s why we’ve put together a team of professional Data Scientists to support you in your journey, meaning you’ll get a response from us within 48 hours maximum.
So, are you ready to embrace the fascinating world of AI?
Come join us, never stop learning, and enjoy AI!
What's included
- 15 hours on-demand video
- 17 articles
- Access on mobile and TV
- Closed captions
- Certificate of completion