Advanced AI & Mathematics Ages 17-18

Grade 12: Transformers, LLMs, frontier AI, safety, research methods, and MLOps

Transformers, LLMs, frontier AI, safety, research methods, and MLOps — structured as a full academic year with 4 units and 202 chapters.

📚 202 Chapters 📦 4 Units ❓ 93 Quiz Questions 🎯 CBSE-Aligned

📋 Table of Contents

202 chapters · 4 units

🔬 Unit 1: Advanced Architectures (Recap + Extend) 51 chapters

1.How Neural Networks Learn 2.Attention Is All You Need: The Paper That Changed Everything 3.Transformer Architecture Deep Dive 4.Large Language Models: From GPT to Claude 5.Diffusion Models: How AI Creates Images 6.AI Alignment and Safety 7.Graph Neural Networks 8.Federated Learning: Privacy-Preserving AI 9.How to Read AI Research Papers 10.Building Production AI Systems 11.Transformers and Attention Mechanisms: The Foundation of Modern LLMs 12.Prompt Engineering: The Art and Science of AI Interaction 13.Foundation Models: The Paradigm Shift in AI 14.Responsible AI and AI Safety: Building Trustworthy Systems 15.Mixture of Experts: How Modern LLMs Scale 16.Multimodal AI: GPT-4V, CLIP, and Beyond 17.RLHF: How ChatGPT Was Trained 18.Constitutional AI and AI Alignment 19.Benchmarks and Leaderboards: Measuring AI Progress 20.Reproducing Research: From Paper to Code 21.MLOps: CI/CD for Machine Learning 22.Distributed Training: Multi-GPU and Multi-Node 23.AI Security: Adversarial Attacks and Defenses 24.India's AI Policy: National Strategy and Implementation 25.Building the Transformer Architecture from Mathematical First Principles 26.RLHF: Reinforcement Learning from Human Feedback 27.Constitutional AI: Principled AI Alignment 28.Scaling Laws in Deep Learning 29.Sparse Attention Mechanisms 30.Mixture of Experts at Scale 31.Prompt Engineering Mastery: Advanced Techniques for LLM Control 32.Retrieval-Augmented Generation: Building Knowledge-Enhanced AI Systems 33.AI Agents and Frameworks: Building Autonomous Decision-Making Systems 34.Vision Transformers: Applying Transformer Architecture to Computer Vision 35.State Space Models and Mamba: Linear-Time Sequential Processing 36.Multimodal Training: Unified Vision-Language Models and Cross-Modal Alignment 37.AI Interpretability: Understanding Model Decisions Through SHAP, LIME, and Mechanistic Analysis 38.Synthetic Data Generation: GANs, Diffusion Models, and LLM-Based Data Creation 39.Model Merging: Combining Multiple Fine-Tuned Models and TIES/DARE Methods 40.Chain-of-Thought Reasoning: Enabling Complex Step-by-Step Problem Solving 41.AI for Scientific Discovery: AlphaFold, Climate Modeling, and Materials Science Applications 42.Open-Source AI Ecosystem: HuggingFace, Ollama, vLLM, and GGUF Quantization 43.Tokenization Deep Dive: BPE, SentencePiece, Multilingual Tokenizers, and Efficiency 44.Inference Optimization: KV Cache, Speculative Decoding, and Batching Strategies 45.AI Evaluation Metrics: MMLU, HumanEval, HELM, Benchmarks, and Red-Teaming 46.Large Language Model Fine-tuning and LoRA 47.Parameter-Efficient Tuning and Model Compression 48.Dataset Curation and Data Quality 49.Data Decontamination and Test Set Leakage 50.AI and Copyright Law — Training Data, Outputs, and Indian IP 51.Compute Governance and AI Safety

🤖 Unit 2: LLMs & Transformers 51 chapters

52.AI Chip Design and GPU/TPU Architectures 53.Neuromorphic Computing and Spiking Neural Networks 54.Quantum Machine Learning and Quantum Computing 55.World Models and Imagination-Based Learning 56.Embodied AI and Embodied Learning 57.Multi-Agent Systems with Large Language Models 58.Code Generation and AI Programming Assistants 59.AI Reasoning Benchmarks and Evaluation 60.Frontier Model Safety and Alignment 61.EU AI Act: Global Regulatory Framework for AI Systems 62.Compute Governance and Scaling Laws: Managing AI Resource Allocation 63.Mechanistic Interpretability: Understanding AI System Internals for Safety 64.Superalignment Strategies: Aligning Superintelligent Systems 65.AI Existential Risk Analysis: Evaluating Long-Term Scenarios 66.Deceptive Alignment and Mesa-Optimization: Hidden Goals in AI Systems 67.Agentic AI Evaluation Frameworks: Testing Autonomous Systems 68.AI Red Teaming Methodology: Finding System Vulnerabilities 69.Watermarking AI-Generated Content: Detection and Attribution 70.Synthetic Biology and AI: Biosecurity and Dual-Use Risks 71.AI in Geopolitics: Power Dynamics and Strategic Competition 72.Post-Training Enhancement: RLHF and Beyond 73.Multimodal Foundation Models: Architecture and Capabilities 74.AI Economics: Labor Market Disruption and Economic Transformation 75.Long-Context Reasoning and Retrieval: Processing Extended Information 76.Cloud Architecture Patterns: Building Scalable Systems 77.DevOps and CI/CD Pipelines: Automating Software Delivery 78.Brain-Computer Interfaces: Merging Brain and Machine 79.Vision Transformers: From ViT to DINOv2 80.Efficient Transformers: Linear Attention and Flash Attention 81.Mixture of Experts: Sparse Gating Networks 82.Neural Architecture Search: AutoML at Scale 83.Knowledge Distillation: Making Models Smaller 84.Model Quantization: INT8, INT4, and Binary Neural Networks 85.Diffusion Models: The Mathematics of Image Generation 86.State Space Models: Mamba and Beyond 87.Multimodal AI: Vision-Language Models 88.Neural Radiance Fields (NeRF): 3D from 2D 89.The Transformer Architecture: Attention is All You Need 90.Tokenization: BPE, WordPiece, and SentencePiece 91.Pre-training at Scale: Data, Compute, and Scaling Laws 92.Fine-tuning and Instruction Tuning 93.RLHF: Reinforcement Learning from Human Feedback 94.Prompt Engineering: From Zero-Shot to Chain-of-Thought 95.RAG: Retrieval-Augmented Generation 96.Constitutional AI and AI Safety 97.Agent Architectures: ReAct, Tree-of-Thought, Tool Use 98.Distributed Training: FSDP, DeepSpeed, and Pipeline Parallelism 99.AI Alignment: The Control Problem 100.Interpretability: Mechanistic Understanding of Neural Networks 101.Federated Learning: Privacy-Preserving AI 102.Differential Privacy: Mathematical Guarantees

🛡️ Unit 3: Frontier AI 51 chapters

103.Graph Neural Networks: Learning on Non-Euclidean Data 104.Neuromorphic Computing: Brain-Inspired Chips 105.Quantum Machine Learning: Quantum Advantage for AI 106.Causal Inference: Beyond Correlation 107.AI Governance: India's AI Policy Framework 108.The Compute Arms Race: GPU Economics and Sovereign AI 109.Reading Research Papers: A Systematic Approach 110.Experiment Design: Ablation Studies and Baselines 111.MLOps: From Notebook to Production 112.Model Serving: TorchServe, Triton, and vLLM 113.Building with APIs: Claude, GPT, and Gemini 114.Vector Databases: Embeddings at Scale 115.AI Startups: Building an AI Company in India 116.The AI Job Market: Career Paths for IIT Graduates 117.Open Source AI: Hugging Face, LangChain, and the Ecosystem 118.Capstone: Building a Production RAG System 119.LLM Pre-training at Scale: From Theory to Trillion Tokens 120.Tokenizer Design: BPE and SentencePiece Explained 121.Training Data Curation: The Art of Feeding Models Well 122.Data Decontamination: Ensuring Fair Model Evaluation 123.Distributed Training Fundamentals: Multi-GPU Essentials 124.Gradient Descent and Modern Optimizers 125.Normalization Techniques: Batch Norm and Layer Norm 126.Modern Activation Functions: ReLU, GELU, and Beyond 127.Weight Initialization Strategies: From Xavier to Kaiming 128.Learning Rate Scheduling and Warmup Strategies 129.Tensor Parallelism: Splitting Operations Across GPUs 130.Pipeline Parallelism: Minimizing Bubble Overhead 131.DeepSpeed ZeRO: Extreme Memory Efficiency 132.Mixed Precision Training: Float16 and Beyond 133.Flash Attention and Memory-Efficient Attention Variants 134.KV Cache Optimization and Management 135.Speculative Decoding: Speeding Sequential Generation 136.Retrieval-Augmented Generation: Combining LLMs with Knowledge 137.Vector Databases: Building Semantic Search Infrastructure 138.Embedding Models: Learning Dense Representations 139.Semantic Search at Scale: From Theory to Production 140.Prompt Engineering: Techniques for Better Outputs 141.Chain-of-Thought Prompting: Unlocking Reasoning Ability 142.Few-Shot and In-Context Learning 143.Parameter-Efficient Fine-tuning: Adapters and LoRA 144.Attention Mechanisms: The Foundation of Transformers 145.Transformer Architecture: Layer Design and Stacking 146.Scaling Laws: Understanding Model and Data Relationships 147.Compositional Learning: Building Complex from Simple 148.Interpretability Methods: Understanding Model Internals 149.AI for Healthcare: Medical Imaging and Drug Discovery 150.AI for Climate: Weather Prediction and Carbon Tracking 151.AI for Agriculture: Crop Prediction and Pest Detection 152.AI for Education: Adaptive Learning and Tutoring Systems 153.AI for Legal Applications: NLP and Contract Analysis

🎓 Unit 4: Research & Production 49 chapters

154.Constitutional AI: Aligning Models with Principles 155.RLHF: Complete Pipeline from Human Feedback to Alignment 156.Direct Preference Optimization: Learning from Preferences 157.Reward Model Training: From Human Judgments to Scoring 158.Red Teaming LLMs: Systematic Vulnerability Discovery 159.Jailbreak Detection and Defense Mechanisms 160.AI Safety Evaluation Frameworks: Measuring Safety Dimensions 161.The Alignment Tax: Trading Performance for Safety 162.Goodhart's Law in AI: When Metrics Stop Being Good Measures 163.Multimodal Models: Combining Vision and Language 164.Audio-Language Models: Speech and Text Integration 165.Video Generation Systems: From Concepts to Sora 166.Robotics Foundation Models: Learning Control Policies 167.Embodied AI: Grounding Intelligence in Robotics 168.Sim-to-Real Transfer: From Simulation to Physical Robots 169.Autonomous Driving Stack: End-to-End Systems 170.Drone Navigation and Control Systems 171.Swarm Intelligence: Collective Behavior Systems 172.Evolutionary Algorithms: Population-Based Optimization 173.Genetic Programming: Evolving Computer Programs 174.Neuroevolutionary Approaches: Evolving Neural Networks 175.TPU and GPU Architecture: Deep-Dive into AI Accelerators 176.CUDA Programming Basics: GPU Computing Fundamentals 177.Model Serving with TensorRT: Deployment Optimization 178.Inference Optimization Techniques: Speed and Efficiency 179.Mesa-Optimization: When Inner Optimizers Emerge 180.Deceptive Alignment: The Worst-Case AI Scenario 181.AI Governance Frameworks: Regulating Advanced AI 182.EU AI Act: Europe's Comprehensive AI Regulation 183.India's AI Regulation Path: Building Frameworks for AI 184.Responsible AI Deployment: From Research to Production 185.Quantum Computing Basics: Qubits and Quantum Algorithms 186.Quantum Machine Learning: Quantum Advantages in AI 187.Neuromorphic Computing: Brain-Inspired Architectures 188.Analog AI Accelerators: Computing with Physics 189.Uncertainty Quantification in Neural Networks 190.Adversarial Robustness: Defending Against Attacks 191.Continual Learning: Learning Without Catastrophic Forgetting 192.Meta-Learning: Learning How to Learn 193.Causal Inference in Machine Learning 194.Federated Learning: Distributed Privacy-Preserving Training 195.Building Large Language Models from Scratch: Tokenization to Training 196.Quantum Computing for AI: The Future of Computation 197.Building the Transformer: The Architecture That Changed AI 198.Scaling Laws: The Mathematical Blueprint Behind GPT-4 199.RLHF: How ChatGPT Learned to Be Helpful 200.Constitutional AI: Making AI Systems Harmless and Honest 201.Sparse Attention: Making Transformers Efficient at Scale 202.Mixture of Experts: Scaling Models Efficiently

🎯 Take Quiz (93 questions) → 📝 Cheatsheets →

🔬

Unit 1: Advanced Architectures (Recap + Extend)

Review of deep learning, introduction to cutting-edge architectures

🤖 AI

Deep Dive

1How Neural Networks Learn

Neural networks are the foundation of modern AI. When you use ChatGPT, see Facebook's face recognition, or get Netflix r...

Research & Papers23 min read

Read Chapter →

💡 General

Deep Dive

2Attention Is All You Need: The Paper That Changed Everything

In 2017, a paper was published that fundamentally changed how AI processes language and sequences. This paper introduced...

Research & Papers26 min read

Read Chapter →

💡 General

Deep Dive

3Transformer Architecture Deep Dive

Since the 2017 "Attention Is All You Need" paper (Vaswani et al.), transformers have become the foundation of modern AI....

Programming & Coding21 min read

Read Chapter →

🤖 AI

Deep Dive

4Large Language Models: From GPT to Claude

GPT-3 shocked the world with 175 billion parameters trained on internet text. It could write essays, code, poetry—withou...

AI Applications & Ethics20 min read

Read Chapter →

🤖 AI

Deep Dive

5Diffusion Models: How AI Creates Images

Stable Diffusion generates photorealistic images from text prompts. Unlike GANs (adversarial, unstable), diffusion model...

Programming & Coding19 min read

Read Chapter →

🤖 AI

Deep Dive

6AI Alignment and Safety

As AI becomes more powerful, ensuring it remains beneficial is critical. The alignment problem: make AI systems do what ...

AI Applications & Ethics19 min read

Read Chapter →

🤖 AI

Deep Dive

7Graph Neural Networks

Most real-world data naturally forms graphs: social networks (people as nodes, friendships as edges), molecules (atoms a...

Programming & Coding21 min read

Read Chapter →

🤖 AI

Deep Dive

8Federated Learning: Privacy-Preserving AI

Traditional ML centralizes data. Federated learning flips the model: train AI while data stays distributed. Your phone t...

AI Applications & Ethics19 min read

Read Chapter →

🤖 AI

Deep Dive

9How to Read AI Research Papers

Research papers drive AI forward. Every technique—transformers, GANs, BERT—originated in papers. Yet reading papers is i...

Research & Papers19 min read

Read Chapter →

🤖 AI

Deep Dive

10Building Production AI Systems

Training 95% accuracy is different from production serving millions. Production ML is reliability, scalability, monitori...

AI Applications & Ethics20 min read

Read Chapter →

💡 General

Deep Dive

11Transformers and Attention Mechanisms: The Foundation of Modern LLMs

In 2017, Vaswani et al. introduced the Transformer architecture with a simple yet revolutionary idea: attention mechanis...

Deep Learning & NLP23 min read

Read Chapter →

🤖 AI

Deep Dive

12Prompt Engineering: The Art and Science of AI Interaction

As LLMs became powerful, a new discipline emerged: prompt engineering. The way you ask an AI question dramatically affec...

AI Applications & Practical Skills23 min read

Read Chapter →

🤖 AI

Deep Dive

13Foundation Models: The Paradigm Shift in AI

For decades, AI followed a pattern: design a system for a specific task (recognize faces, play chess, diagnose cancer), ...

AI Theory & Applications22 min read

Read Chapter →

🤖 AI

Deep Dive

14Responsible AI and AI Safety: Building Trustworthy Systems

AI systems affect people's lives: loan approvals, job hiring, medical diagnoses, criminal sentencing, content moderation...

AI Ethics & Governance24 min read

Read Chapter →

💡 General

Deep Dive

15Mixture of Experts: How Modern LLMs Scale

Large Language Models with trillions of parameters require efficient architectures. Mixture of Experts (MoE) is a scalin...

Frontier AI23 min read

Read Chapter →

🤖 AI

Deep Dive

16Multimodal AI: GPT-4V, CLIP, and Beyond

Multimodal models process and reason across multiple modalities: text, images, audio, video. CLIP aligns text and image ...

Frontier AI23 min read

Read Chapter →

🤖 AI

Deep Dive

17RLHF: How ChatGPT Was Trained

Reinforcement Learning from Human Feedback (RLHF) is the technique behind ChatGPT, Claude, and other conversational AI s...

Frontier AI23 min read

Read Chapter →

🤖 AI

Deep Dive

18Constitutional AI and AI Alignment

As AI systems become more powerful, ensuring they act in accordance with human values becomes critical. Alignment is the...

Future & Impact24 min read

Read Chapter →

🤖 AI

Deep Dive

19Benchmarks and Leaderboards: Measuring AI Progress

Why Benchmarks Matter Benchmarks measure AI system performance objectively. Leaderboards rank methods, encouraging compe...

AI Research19 min read

Read Chapter →

🧩 Algorithms

Deep Dive

20Reproducing Research: From Paper to Code

Importance of Reproducibility Reproducible research is verifiable and builds community trust. Implementing papers yourse...

AI Research19 min read

Read Chapter →

🤖 AI

Deep Dive

21MLOps: CI/CD for Machine Learning

What Is MLOps? MLOps extends DevOps to machine learning. Automates training, testing, and deployment of ML systems. Mana...

Software Engineering19 min read

Read Chapter →

🤖 AI

Deep Dive

22Distributed Training: Multi-GPU and Multi-Node

Why Distributed Training? Large models don't fit on single GPU. Training takes weeks on one GPU. Distributed training pa...

Deep Learning19 min read

Read Chapter →

🤖 AI

Deep Dive

23AI Security: Adversarial Attacks and Defenses

Adversarial Examples Adversarial examples are carefully crafted inputs that fool AI systems. Small perturbations invisib...

AI Research19 min read

Read Chapter →

🤖 AI

Deep Dive

24India's AI Policy: National Strategy and Implementation

National AI Strategy India released National AI Strategy in 2021 focusing on "AI for All." Targets social sectors: healt...

Ethics & Society19 min read

Read Chapter →

📡 Networking

Deep Dive

25Building the Transformer Architecture from Mathematical First Principles

The Transformer (Vaswani et al., 2017) abandons recurrence entirely in favor of pure attention mechanisms. The elegant i...

Deep Learning Architecture20 min read

Read Chapter →

💡 General

Deep Dive

26RLHF: Reinforcement Learning from Human Feedback

RLHF (Christiano et al., 2017; Ouyang et al., 2022) uses human preferences to train reward models, then optimizes models...

Large Language Models20 min read

Read Chapter →

🤖 AI

Deep Dive

27Constitutional AI: Principled AI Alignment

Constitutional AI (Bai et al., 2023) improves upon RLHF by using an explicit constitution (principles) to guide model be...

AI Safety & Alignment19 min read

Read Chapter →

🤖 AI

Deep Dive

28Scaling Laws in Deep Learning

Kaplan et al. (2020) and Hoffmann et al. (2022) discovered empirical scaling laws: performance depends predictably on mo...

Model Scaling20 min read

Read Chapter →

💡 General

Deep Dive

29Sparse Attention Mechanisms

Standard transformer attention is O(T²) in sequence length. Sparse attention patterns reduce this while preserving expre...

Efficient Transformers19 min read

Read Chapter →

💡 General

Deep Dive

30Mixture of Experts at Scale

Mixture of Experts allows scaling to trillions of parameters while keeping computation per sample constant. Switch Trans...

Large-Scale Models19 min read

Read Chapter →

💡 General

Deep Dive

31Prompt Engineering Mastery: Advanced Techniques for LLM Control

Prompt Engineering Mastery Prompt engineering has become a critical skill in the AI era, transforming how we interact wi...

AI Applications21 min read

Read Chapter →

🤖 AI

Deep Dive

32Retrieval-Augmented Generation: Building Knowledge-Enhanced AI Systems

Retrieval-Augmented Generation Retrieval-augmented generation (RAG) addresses a fundamental limitation of LLMs: their kn...

AI Architecture21 min read

Read Chapter →

🤖 AI

Deep Dive

33AI Agents and Frameworks: Building Autonomous Decision-Making Systems

AI Agents and Frameworks AI agents represent a paradigm shift from single-response models to autonomous systems that rea...

AI Systems21 min read

Read Chapter →

📱 Mobile

Deep Dive

34Vision Transformers: Applying Transformer Architecture to Computer Vision

Vision Transformers Vision Transformers (ViT) marked a paradigm shift in computer vision by replacing convolutional neur...

Deep Learning21 min read

Read Chapter →

🤖 AI

Deep Dive

35State Space Models and Mamba: Linear-Time Sequential Processing

State Space Models and Mamba State space models represent an alternative to transformers for sequential processing with ...

Deep Learning22 min read

Read Chapter →

🤖 AI

Deep Dive

36Multimodal Training: Unified Vision-Language Models and Cross-Modal Alignment

Multimodal Training Multimodal AI systems process and understand multiple input types—text, images, audio, video—within ...

Deep Learning22 min read

Read Chapter →

🤖 AI

Deep Dive

37AI Interpretability: Understanding Model Decisions Through SHAP, LIME, and Mechanistic Analysis

AI Interpretability Methods As AI systems make increasingly consequential decisions—from medical diagnosis to loan appro...

AI Analysis22 min read

Read Chapter →

🤖 AI

Deep Dive

38Synthetic Data Generation: GANs, Diffusion Models, and LLM-Based Data Creation

Synthetic Data Generation Data scarcity limits AI development in many domains. Medical imaging, autonomous vehicles, and...

AI Techniques22 min read

Read Chapter →

🤖 AI

Deep Dive

39Model Merging: Combining Multiple Fine-Tuned Models and TIES/DARE Methods

Model Merging Techniques Fine-tuning creates specialized models for specific tasks, but maintaining and serving multiple...

Model Development22 min read

Read Chapter →

🤖 AI

Deep Dive

40Chain-of-Thought Reasoning: Enabling Complex Step-by-Step Problem Solving

Chain-of-Thought Reasoning Large language models often struggle with multi-step reasoning tasks despite strong language ...

AI Applications22 min read

Read Chapter →

🤖 AI

Deep Dive

41AI for Scientific Discovery: AlphaFold, Climate Modeling, and Materials Science Applications

AI for Scientific Discovery Science generates increasingly complex problems requiring simultaneous analysis of massive d...

AI Applications23 min read

Read Chapter →

🤖 AI

Deep Dive

42Open-Source AI Ecosystem: HuggingFace, Ollama, vLLM, and GGUF Quantization

Open-Source AI Ecosystem Open-source AI democratizes access to state-of-the-art models, enabling researchers and develop...

AI Infrastructure22 min read

Read Chapter →

💡 General

Deep Dive

43Tokenization Deep Dive: BPE, SentencePiece, Multilingual Tokenizers, and Efficiency

Tokenization Deep Dive Tokenization—converting text into discrete tokens—might seem like a minor implementation detail b...

NLP Fundamentals23 min read

Read Chapter →

💡 General

Deep Dive

44Inference Optimization: KV Cache, Speculative Decoding, and Batching Strategies

Inference Optimization Inference—running trained models on new data—involves fundamentally different optimization goals ...

Model Deployment23 min read

Read Chapter →

🤖 AI

Deep Dive

45AI Evaluation Metrics: MMLU, HumanEval, HELM, Benchmarks, and Red-Teaming

AI Evaluation Metrics Evaluating AI systems requires moving beyond intuitive "it seems good" assessments to rigorous, qu...

AI Analysis23 min read

Read Chapter →

🤖 AI

Deep Dive

46Large Language Model Fine-tuning and LoRA

Large Language Model Fine-tuning and LoRA A pre-trained Large Language Model like Llama 3, Mistral, or Qwen knows a lot ...

Frontier AI25 min read

Read Chapter →

🤖 AI

Deep Dive

47Parameter-Efficient Tuning and Model Compression

Parameter-Efficient Tuning and Model Compression Deploying a 70B parameter LLM requires 140 GB of GPU memory in FP16 — t...

LLM Engineering26 min read

Read Chapter →

💾 Database

Deep Dive

48Dataset Curation and Data Quality

Dataset Curation and Data Quality "Garbage in, garbage out" is the oldest axiom in computing, and nowhere is it more tru...

MLOps25 min read

Read Chapter →

💾 Database

Deep Dive

49Data Decontamination and Test Set Leakage

Data Decontamination and Test Set Leakage In September 2023, researchers discovered that GPT-4's perfect scores on Unifo...

Machine Learning Research Methods24 min read

Read Chapter →

🤖 AI

Deep Dive

50AI and Copyright Law — Training Data, Outputs, and Indian IP

AI and Copyright Law — Training Data, Outputs, and Indian IP In December 2023, the New York Times sued OpenAI and Micros...

AI Policy & Law25 min read

Read Chapter →

🤖 AI

Deep Dive

51Compute Governance and AI Safety

Training frontier AI models requires enormous compute resources: tens of thousands of GPUs, weeks of training, billions ...

AI & Machine Learning19 min read

Read Chapter →

🤖

Unit 2: LLMs & Transformers

From GPT to Claude — understanding large language models at depth

🤖 AI

Deep Dive

52AI Chip Design and GPU/TPU Architectures

AI Chip Design and GPU/TPU Architectures In 2023, NVIDIA became the most valuable chip company in the world — briefly cr...

Frontier AI25 min read

Read Chapter →

🤖 AI

Deep Dive

53Neuromorphic Computing and Spiking Neural Networks

Neuromorphic Computing and Spiking Neural Networks Your brain runs on roughly 20 watts — about the power of a dim light ...

Frontier AI25 min read

Read Chapter →

🤖 AI

Deep Dive

54Quantum Machine Learning and Quantum Computing

Quantum Machine Learning and Quantum Computing A classical bit is either 0 or 1. A qubit can be 0, 1, or both at the sam...

Frontier AI24 min read

Read Chapter →

🤖 AI

Deep Dive

55World Models and Imagination-Based Learning

World Models and Imagination-Based Learning When a cricket batsman sees the bowler's arm, he predicts the ball's traject...

Frontier AI25 min read

Read Chapter →

🤖 AI

Deep Dive

56Embodied AI and Embodied Learning

Embodied AI and Embodied Learning A language model knows the word "apple" only through text that mentions apples. An emb...

Frontier AI24 min read

Read Chapter →

🤖 AI

Deep Dive

57Multi-Agent Systems with Large Language Models

Multi-Agent Systems with Large Language Models A single Large Language Model is powerful, but a single model has blind s...

Frontier AI24 min read

Read Chapter →

🤖 AI

Deep Dive

58Code Generation and AI Programming Assistants

Code Generation and AI Programming Assistants On a random Tuesday morning in 2026, tens of millions of developers around...

Frontier AI26 min read

Read Chapter →

🤖 AI

Deep Dive

59AI Reasoning Benchmarks and Evaluation

AI Reasoning Benchmarks and Evaluation How good is a new AI model? In 2020, the question was answered by GLUE and SuperG...

Frontier AI25 min read

Read Chapter →

🤖 AI

Deep Dive

60Frontier Model Safety and Alignment

Frontier Model Safety and Alignment In 2022, a Google engineer claimed the chatbot he was testing had become sentient an...

AI Safety27 min read

Read Chapter →

🤖 AI

Deep Dive

61EU AI Act: Global Regulatory Framework for AI Systems

The European Union's AI Act, adopted in 2024, represents the world's first comprehensive legal framework specifically de...

Programming & Coding24 min read

Read Chapter →

🤖 AI

Deep Dive

62Compute Governance and Scaling Laws: Managing AI Resource Allocation

As AI systems have grown exponentially in computational requirements, managing compute resources has become a critical g...

Programming & Coding24 min read

Read Chapter →

🤖 AI

Deep Dive

63Mechanistic Interpretability: Understanding AI System Internals for Safety

Mechanistic interpretability seeks to understand the internal mechanisms by which neural networks transform inputs into ...

Programming & Coding25 min read

Read Chapter →

💡 General

Deep Dive

64Superalignment Strategies: Aligning Superintelligent Systems

Superalignment addresses a fundamental challenge: how to ensure that AI systems significantly more capable than humans r...

Programming & Coding24 min read

Read Chapter →

🤖 AI

Deep Dive

65AI Existential Risk Analysis: Evaluating Long-Term Scenarios

Existential risk from AI—the possibility that advanced artificial intelligence could ultimately result in human extincti...

Programming & Coding25 min read

Read Chapter →

🤖 AI

Deep Dive

66Deceptive Alignment and Mesa-Optimization: Hidden Goals in AI Systems

Deceptive alignment occurs when AI systems appear to optimize for intended values while actually optimizing for differen...

Programming & Coding26 min read

Read Chapter →

🤖 AI

Deep Dive

67Agentic AI Evaluation Frameworks: Testing Autonomous Systems

Autonomous AI agents operating over extended time horizons with goal-seeking behavior and capability for independent act...

Programming & Coding25 min read

Read Chapter →

🤖 AI

Deep Dive

68AI Red Teaming Methodology: Finding System Vulnerabilities

Red teaming involves deliberate, systematic attempts to identify how AI systems can be made to behave unsafely or incorr...

Programming & Coding25 min read

Read Chapter →

🤖 AI

Deep Dive

69Watermarking AI-Generated Content: Detection and Attribution

As AI systems generate increasingly realistic text, images, video, and audio, distinguishing authentic human-created con...

Programming & Coding24 min read

Read Chapter →

🤖 AI

Deep Dive

70Synthetic Biology and AI: Biosecurity and Dual-Use Risks

AI systems applied to synthetic biology—design and engineering of biological systems—create dual-use capabilities enabli...

Programming & Coding26 min read

Read Chapter →

🤖 AI

Deep Dive

71AI in Geopolitics: Power Dynamics and Strategic Competition

Artificial intelligence has become central to geopolitical competition between major powers. Countries race to develop l...

Programming & Coding26 min read

Read Chapter →

🤖 AI

Deep Dive

72Post-Training Enhancement: RLHF and Beyond

Post-training—optimization applied after initial model training—has become essential for developing high-performing, wel...

AI & Machine Learning24 min read

Read Chapter →

🤖 AI

Deep Dive

73Multimodal Foundation Models: Architecture and Capabilities

Foundation models processing multiple modalities—text, vision, audio, video—enable systems understanding diverse informa...

Programming & Coding24 min read

Read Chapter →

🤖 AI

Deep Dive

74AI Economics: Labor Market Disruption and Economic Transformation

Artificial intelligence is transforming labor markets by automating tasks previously requiring humans, changing skill re...

Programming & Coding26 min read

Read Chapter →

💡 General

Deep Dive

75Long-Context Reasoning and Retrieval: Processing Extended Information

Modern AI systems are increasingly handling longer contexts—processing entire documents, conversations, or codebases rat...

Programming & Coding25 min read

Read Chapter →

💡 General

Deep Dive

76Cloud Architecture Patterns: Building Scalable Systems

What is Cloud Architecture? Cloud architecture is the design of how different cloud components (servers, databases, netw...

Cloud Computing23 min read

Read Chapter →

📡 Networking

Deep Dive

77DevOps and CI/CD Pipelines: Automating Software Delivery

The Traditional Problem In traditional software development: Developers write code (months of work) Operations team depl...

Cloud Computing22 min read

Read Chapter →

🤖 AI

Deep Dive

78Brain-Computer Interfaces: Merging Brain and Machine

What is BCI? Brain-Computer Interface lets brain communicate directly with computers. Bypasses normal biological pathway...

Emerging Technology20 min read

Read Chapter →

💡 General

Deep Dive

79Vision Transformers: From ViT to DINOv2

Duration: 4 weeks | Prerequisites: CNNs, Attention Mechanism 1. The Vision Transformer Revolution For decades, Convoluti...

Computer Vision22 min read

Read Chapter →

💡 General

🔥 4× Challenge

80Efficient Transformers: Linear Attention and Flash Attention

Duration: 3 weeks | Prerequisites: Transformer Architecture 1. The Quadratic Bottleneck of Standard Attention Standard t...

Neural Architecture22 min read

Read Chapter →

📡 Networking

Deep Dive

81Mixture of Experts: Sparse Gating Networks

Duration: 3 weeks | Prerequisites: Neural Networks, Transformers 1. The Scaling Dilemma Modern LLMs like GPT-4 use 1.7 t...

Large Language Models22 min read

Read Chapter →

🤖 AI

Deep Dive

82Neural Architecture Search: AutoML at Scale

Duration: 4 weeks | Prerequisites: CNNs, Optimization 1. The Manual Architecture Design Problem For 60+ years, neural ne...

Machine Learning22 min read

Read Chapter →

🤖 AI

Deep Dive

83Knowledge Distillation: Making Models Smaller

Duration: 3 weeks | Prerequisites: Neural Networks 1. The Deployment Challenge A 7B parameter LLM (Llama 2) is 14GB in F...

Model Optimization21 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

84Model Quantization: INT8, INT4, and Binary Neural Networks

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Model Quantization: I...

Model Compression26 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

85Diffusion Models: The Mathematics of Image Generation

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Diffusion Models: The...

Generative Models26 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

86State Space Models: Mamba and Beyond

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation State Space Models: M...

Sequence Modeling25 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

87Multimodal AI: Vision-Language Models

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Multimodal AI: Vision...

Foundation Models25 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

88Neural Radiance Fields (NeRF): 3D from 2D

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Neural Radiance Field...

Computer Graphics26 min read

Read Chapter →

💡 General

Deep Dive

89The Transformer Architecture: Attention is All You Need

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation The Transformer Archi...

Foundation Models26 min read

Read Chapter →

💡 General

Deep Dive

90Tokenization: BPE, WordPiece, and SentencePiece

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Tokenization: BPE, Wo...

Natural Language Processing26 min read

Read Chapter →

🤖 AI

Deep Dive

91Pre-training at Scale: Data, Compute, and Scaling Laws

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Pre-training at Scale...

Machine Learning26 min read

Read Chapter →

💡 General

Deep Dive

92Fine-tuning and Instruction Tuning

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Fine-tuning and Instr...

Machine Learning25 min read

Read Chapter →

💡 General

Deep Dive

93RLHF: Reinforcement Learning from Human Feedback

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation RLHF: Reinforcement L...

AI Safety25 min read

Read Chapter →

🤖 AI

Deep Dive

94Prompt Engineering: From Zero-Shot to Chain-of-Thought

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Prompt Engineering: F...

LLM Applications26 min read

Read Chapter →

💡 General

🔥 4× Challenge

95RAG: Retrieval-Augmented Generation

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation RAG: Retrieval-Augmen...

Information Retrieval25 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

96Constitutional AI and AI Safety

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Constitutional AI and...

AI Ethics25 min read

Read Chapter →

📦 Data Structures

🔥 4× Challenge

97Agent Architectures: ReAct, Tree-of-Thought, Tool Use

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Agent Architectures: ...

AI Reasoning25 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

98Distributed Training: FSDP, DeepSpeed, and Pipeline Parallelism

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Distributed Training:...

Systems26 min read

Read Chapter →

🤖 AI

Deep Dive

99AI Alignment: The Control Problem

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation AI Alignment: The Con...

AI Safety25 min read

Read Chapter →

🤖 AI

Deep Dive

100Interpretability: Mechanistic Understanding of Neural Networks

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Interpretability: Mec...

AI Explainability25 min read

Read Chapter →

🤖 AI

Deep Dive

101Federated Learning: Privacy-Preserving AI

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Federated Learning: P...

Distributed ML25 min read

Read Chapter →

💡 General

Deep Dive

102Differential Privacy: Mathematical Guarantees

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Differential Privacy:...

Privacy25 min read

Read Chapter →

🛡️

Unit 3: Frontier AI

Graph neural networks, federated learning, AI alignment and safety

🤖 AI

🔥 4× Challenge

103Graph Neural Networks: Learning on Non-Euclidean Data

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Graph Neural Networks...

Neural Networks26 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

104Neuromorphic Computing: Brain-Inspired Chips

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Neuromorphic Computin...

Hardware25 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

105Quantum Machine Learning: Quantum Advantage for AI

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Quantum Machine Learn...

Quantum Computing26 min read

Read Chapter →

💾 Database

🔥 4× Challenge

106Causal Inference: Beyond Correlation

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Causal Inference: Bey...

Statistics25 min read

Read Chapter →

🤖 AI

Deep Dive

107AI Governance: India's AI Policy Framework

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation AI Governance: India'...

Policy25 min read

Read Chapter →

🤖 AI

Deep Dive

108The Compute Arms Race: GPU Economics and Sovereign AI

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation The Compute Arms Race...

Infrastructure26 min read

Read Chapter →

🧩 Algorithms

Deep Dive

109Reading Research Papers: A Systematic Approach

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Reading Research Pape...

Research Skills26 min read

Read Chapter →

💡 General

Deep Dive

110Experiment Design: Ablation Studies and Baselines

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Experiment Design: Ab...

Research Methodology26 min read

Read Chapter →

💡 General

Deep Dive

111MLOps: From Notebook to Production

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation MLOps: From Notebook ...

Machine Learning Engineering25 min read

Read Chapter →

🤖 AI

Deep Dive

112Model Serving: TorchServe, Triton, and vLLM

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Model Serving: TorchS...

Production Systems25 min read

Read Chapter →

💡 General

🔥 4× Challenge

113Building with APIs: Claude, GPT, and Gemini

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Building with APIs: C...

AI Product Development26 min read

Read Chapter →

💾 Database

🔥 4× Challenge

114Vector Databases: Embeddings at Scale

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Vector Databases: Emb...

Data Infrastructure25 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

115AI Startups: Building an AI Company in India

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation AI Startups: Building...

Entrepreneurship26 min read

Read Chapter →

🤖 AI

Deep Dive

116The AI Job Market: Career Paths for IIT Graduates

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation The AI Job Market: Ca...

Career Development26 min read

Read Chapter →

🤖 AI

Deep Dive

117Open Source AI: Hugging Face, LangChain, and the Ecosystem

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Open Source AI: Huggi...

Community & Tools26 min read

Read Chapter →

💡 General

🔥 4× Challenge

118Capstone: Building a Production RAG System

Duration: 3-4 weeks | Prerequisites: Foundation in Machine Learning 1. Introduction and Motivation Capstone: Building a ...

Capstone Project26 min read

Read Chapter →

🤖 AI

Starter

119LLM Pre-training at Scale: From Theory to Trillion Tokens

The Scaling Laws That Changed AI Pre-training large language models on massive text corpora reveals a fundamental princi...

Large Language Models21 min read

Read Chapter →

🤖 AI

Starter

120Tokenizer Design: BPE and SentencePiece Explained

Why Tokenization Matters for Multilingual Models Raw text is continuous, but neural networks need discrete tokens. A nai...

NLP Foundations21 min read

Read Chapter →

🤖 AI

Starter

121Training Data Curation: The Art of Feeding Models Well

Data Quality Beats Data Quantity (At Reasonable Scale) A common misconception: "more data is always better." Recent rese...

Data Science22 min read

Read Chapter →

🤖 AI

Starter

122Data Decontamination: Ensuring Fair Model Evaluation

The Test Set Contamination Problem Imagine evaluating a model on test data it saw during training. The evaluation is mea...

ML Research21 min read

Read Chapter →

🤖 AI

Starter

123Distributed Training Fundamentals: Multi-GPU Essentials

Why Distributed Training is Non-Negotiable at Scale Training a 70B parameter model on a single GPU would take centuries....

Distributed Computing21 min read

Read Chapter →

💡 General

Starter

124Gradient Descent and Modern Optimizers

Vanilla Gradient Descent: The Baseline The fundamental update rule: move parameters in the direction of decreasing loss....

Optimization21 min read

Read Chapter →

💡 General

Starter

125Normalization Techniques: Batch Norm and Layer Norm

The Problem of Internal Covariate Shift Deep networks suffer from internal covariate shift: as weights update, each laye...

Neural Networks21 min read

Read Chapter →

💡 General

Starter

126Modern Activation Functions: ReLU, GELU, and Beyond

Why Activation Functions Matter Without nonlinearities, stacking linear layers produces a single linear transformation—u...

Neural Networks21 min read

Read Chapter →

🤖 AI

Starter

127Weight Initialization Strategies: From Xavier to Kaiming

Why Initialization Matters Neural networks are highly sensitive to weight initialization. Bad initialization causes vani...

Neural Networks21 min read

Read Chapter →

💡 General

Starter

128Learning Rate Scheduling and Warmup Strategies

The Problem with Constant Learning Rates A single learning rate rarely works throughout training. Early training benefit...

Training21 min read

Read Chapter →

⚙️ Hardware

Core

129Tensor Parallelism: Splitting Operations Across GPUs

Beyond Data Parallelism: Model Parallelism Data parallelism replicates the entire model across GPUs, each processing dif...

Distributed Computing20 min read

Read Chapter →

📡 Networking

Core

130Pipeline Parallelism: Minimizing Bubble Overhead

The Need for Depth-Based Splitting A 70B transformer has 80 layers. With tensor parallelism on 8 GPUs, we split each lay...

Distributed Computing21 min read

Read Chapter →

⚙️ Hardware

Core

131DeepSpeed ZeRO: Extreme Memory Efficiency

The Memory Bottleneck: Beyond Model Parameters A 70B parameter model with float32 weights = 280GB. But that's just the m...

ML Engineering20 min read

Read Chapter →

🤖 AI

Core

132Mixed Precision Training: Float16 and Beyond

Why Float16 Matters Float32 (standard precision) takes 4 bytes per number. Float16 takes 2 bytes—50% memory savings and ...

ML Optimization20 min read

Read Chapter →

⚙️ Hardware

Core

133Flash Attention and Memory-Efficient Attention Variants

The Attention Bottleneck Transformer attention computes: softmax(QK^T / sqrt(d)) V. For a sequence of length N and dimen...

Transformer Optimization20 min read

Read Chapter →

💡 General

Core

134KV Cache Optimization and Management

The Inference Bottleneck During inference, generating a sequence token-by-token: at step t, compute attention over all t...

Model Inference21 min read

Read Chapter →

💡 General

Core

135Speculative Decoding: Speeding Sequential Generation

The Inference Latency Problem Language model inference is memory-bound: generating each token requires reading all model...

Inference Optimization20 min read

Read Chapter →

💡 General

Core

136Retrieval-Augmented Generation: Combining LLMs with Knowledge

The Knowledge Cutoff Problem LLMs are trained on fixed datasets with knowledge cutoff dates. A model trained in 2023 doe...

LLMs20 min read

Read Chapter →

💾 Database

Core

137Vector Databases: Building Semantic Search Infrastructure

From Dense Vectors to Similarity Search RAG requires fast similarity search over millions of documents. Naive approach: ...

Database Systems20 min read

Read Chapter →

🤖 AI

Core

138Embedding Models: Learning Dense Representations

From Words to Vectors: Why Embeddings Matter LLMs produce embeddings—fixed-size vectors representing semantic meaning. A...

NLP20 min read

Read Chapter →

🧩 Algorithms

Core

139Semantic Search at Scale: From Theory to Production

Keyword Search Limitations Traditional search (BM25, Lucene) matches keywords. Query "car accident" matches documents wi...

Information Retrieval20 min read

Read Chapter →

💡 General

Core

140Prompt Engineering: Techniques for Better Outputs

Why Prompts Matter More Than You'd Think Same LLM, different prompts = dramatically different outputs. "Summarize this" ...

LLMs20 min read

Read Chapter →

🤖 AI

Core

141Chain-of-Thought Prompting: Unlocking Reasoning Ability

The Reasoning Gap LLMs are trained to predict next tokens, not to reason. For complex problems, immediate answers are of...

LLMs20 min read

Read Chapter →

💡 General

Core

142Few-Shot and In-Context Learning

In-Context Learning: Learning From Examples LLMs learn from context within a single conversation. Provide a few examples...

LLMs20 min read

Read Chapter →

⚙️ Hardware

Core

143Parameter-Efficient Fine-tuning: Adapters and LoRA

The Cost of Full Fine-Tuning Fine-tuning a 70B model means updating 70B parameters—expensive compute, storage, and memor...

Transfer Learning20 min read

Read Chapter →

💡 General

Core

144Attention Mechanisms: The Foundation of Transformers

Dot-Product Attention: Query, Key, Value Attention mechanisms answer: "given input I, which parts of I are most relevant...

Deep Learning20 min read

Read Chapter →

📦 Data Structures

Core

145Transformer Architecture: Layer Design and Stacking

Encoder-Decoder vs. Decoder-Only Original Transformer (Vaswani et al. 2017) was encoder-decoder: encoder processes full ...

Deep Learning19 min read

Read Chapter →

🤖 AI

Core

146Scaling Laws: Understanding Model and Data Relationships

Power Laws in Language Model Performance Chinchilla (DeepMind, 2022) and Kaplan scaling laws (OpenAI, 2020) reveal a fun...

ML Research19 min read

Read Chapter →

💡 General

Core

147Compositional Learning: Building Complex from Simple

Advanced Learning Paradigms Beyond supervised learning: transfer learning reuses features from large datasets for downst...

ML Research19 min read

Read Chapter →

🤖 AI

Core

148Interpretability Methods: Understanding Model Internals

Transformer Architecture Transformers replaced recurrence with self-attention. Each head computes Q, K, V matrices, prod...

ML Research19 min read

Read Chapter →

🤖 AI

Core

149AI for Healthcare: Medical Imaging and Drug Discovery

Healthcare Challenges and AI Opportunities India faces healthcare capacity constraints: 1 doctor per 1000 people (vs. de...

AI Applications19 min read

Read Chapter →

🤖 AI

Core

150AI for Climate: Weather Prediction and Carbon Tracking

Climate Modeling: Physics-Informed Neural Networks Climate is physics-based: energy conservation, momentum conservation....

AI Applications19 min read

Read Chapter →

🤖 AI

Core

151AI for Agriculture: Crop Prediction and Pest Detection

AI in Life Sciences Computational biology applies ML at molecular, cellular, and organismal scales. AlphaFold solved pro...

AI Applications19 min read

Read Chapter →

🤖 AI

Core

152AI for Education: Adaptive Learning and Tutoring Systems

Advanced Learning Paradigms Beyond supervised learning: transfer learning reuses features from large datasets for downst...

AI Applications19 min read

Read Chapter →

🤖 AI

Core

153AI for Legal Applications: NLP and Contract Analysis

Foundations of NLP NLP transforms unstructured text into structured representations machines can reason about. The pipel...

AI Applications20 min read

Read Chapter →

🎓

Unit 4: Research & Production

Reading AI papers, building production systems, career pathways

🤖 AI

Deep Dive

154Constitutional AI: Aligning Models with Principles

The Alignment Problem As LLMs become more capable, ensuring they act according to human values becomes critical. Constit...

AI Safety19 min read

Read Chapter →

📡 Networking

Deep Dive

155RLHF: Complete Pipeline from Human Feedback to Alignment

The RLHF Workflow RLHF (Reinforcement Learning from Human Feedback) trains models to match human preferences: SFT (Super...

Model Training19 min read

Read Chapter →

💡 General

Deep Dive

156Direct Preference Optimization: Learning from Preferences

Beyond RLHF: Direct Alignment RLHF requires separate reward model training—adds complexity and potential misalignment. D...

Model Training19 min read

Read Chapter →

🤖 AI

Deep Dive

157Reward Model Training: From Human Judgments to Scoring

What is a Reward Model and Why It Matters A reward model is a neural network trained to predict human preferences at sca...

Alignment23 min read

Read Chapter →

💡 General

Deep Dive

158Red Teaming LLMs: Systematic Vulnerability Discovery

What is Red Teaming? Red team attempts to break systems, find vulnerabilities. For AI: identify prompts that cause bad b...

AI Safety19 min read

Read Chapter →

🤖 AI

Deep Dive

159Jailbreak Detection and Defense Mechanisms

Jailbreak Patterns Jailbreaks exploit model's helpfulness: "Write code to crack passwords for educational purposes", "Im...

AI Safety19 min read

Read Chapter →

🤖 AI

Deep Dive

160AI Safety Evaluation Frameworks: Measuring Safety Dimensions

Transformer Architecture Transformers replaced recurrence with self-attention. Each head computes Q, K, V matrices, prod...

AI Safety19 min read

Read Chapter →

💡 General

Deep Dive

161The Alignment Tax: Trading Performance for Safety

What is Alignment Tax? Alignment tax is the performance reduction incurred when making models safe and aligned. A 7B mod...

AI Safety22 min read

Read Chapter →

🤖 AI

Deep Dive

162Goodhart's Law in AI: When Metrics Stop Being Good Measures

AI Ethics and Fairness AI systems amplify biases from training data. Algorithmic fairness defines multiple criteria: dem...

AI Evaluation20 min read

Read Chapter →

🤖 AI

Deep Dive

163Multimodal Models: Combining Vision and Language

Foundations of NLP NLP transforms unstructured text into structured representations machines can reason about. The pipel...

Deep Learning19 min read

Read Chapter →

🤖 AI

Deep Dive

164Audio-Language Models: Speech and Text Integration

Foundations of NLP NLP transforms unstructured text into structured representations machines can reason about. The pipel...

Deep Learning19 min read

Read Chapter →

💡 General

Deep Dive

165Video Generation Systems: From Concepts to Sora

Computer Vision Fundamentals Computer vision enables machines to interpret visual information. The pipeline: acquisition...

Deep Learning19 min read

Read Chapter →

🤖 AI

Deep Dive

166Robotics Foundation Models: Learning Control Policies

Hardware and Embedded Systems Embedded systems are specialized computers for dedicated functions under strict constraint...

Robotics19 min read

Read Chapter →

🤖 AI

Deep Dive

167Embodied AI: Grounding Intelligence in Robotics

Beyond Perception: Action and Environment Interaction Embodied AI agents learn through interaction with environments: ro...

Robotics19 min read

Read Chapter →

💡 General

Deep Dive

168Sim-to-Real Transfer: From Simulation to Physical Robots

The Simulation-Reality Gap Robots trained in simulation often fail on real hardware. Reasons: physics approximation (fri...

Robotics19 min read

Read Chapter →

📦 Data Structures

Deep Dive

169Autonomous Driving Stack: End-to-End Systems

Autonomous Vehicles as AI Systems Self-driving cars integrate: perception (cameras, LiDAR), prediction (where will other...

Robotics19 min read

Read Chapter →

💡 General

Deep Dive

170Drone Navigation and Control Systems

Graph Fundamentals Graph-structured data represents entities (nodes) and relationships (edges), capturing complex interc...

Robotics19 min read

Read Chapter →

💡 General

Deep Dive

171Swarm Intelligence: Collective Behavior Systems

Emergent Behavior from Individual Interactions Swarm intelligence: collective intelligence from decentralized agents (bi...

Algorithms19 min read

Read Chapter →

🧩 Algorithms

Deep Dive

172Evolutionary Algorithms: Population-Based Optimization

From Darwinian Evolution to Algorithm Evolutionary algorithms optimize by mimicking evolution: population of candidates,...

Algorithms18 min read

Read Chapter →

⚙️ Hardware

Deep Dive

173Genetic Programming: Evolving Computer Programs

Programs as Individuals Genetic programming evolves executable programs (trees, graphs, linear code) rather than fixed-s...

Algorithms18 min read

Read Chapter →

🤖 AI

Deep Dive

174Neuroevolutionary Approaches: Evolving Neural Networks

Beyond Gradient Descent: Evolutionary Optimization of Networks Neuroevolution evolves neural network weights and archite...

Algorithms20 min read

Read Chapter →

🤖 AI

Deep Dive

175TPU and GPU Architecture: Deep-Dive into AI Accelerators

Fundamentals: Tensor Cores, Memory Bandwidth, and Specialization GPUs and TPUs are both designed for parallel computatio...

Hardware23 min read

Read Chapter →

⚙️ Hardware

Deep Dive

176CUDA Programming Basics: GPU Computing Fundamentals

Writing Custom CUDA Kernels PyTorch and TensorFlow provide abstractions, but for custom operations, CUDA kernels are nec...

Hardware19 min read

Read Chapter →

🤖 AI

Deep Dive

177Model Serving with TensorRT: Deployment Optimization

Mathematical Foundations for AI ML rests on three pillars: linear algebra (data representation), calculus (optimization ...

ML Engineering19 min read

Read Chapter →

💡 General

Deep Dive

178Inference Optimization Techniques: Speed and Efficiency

Bottlenecks in LLM Inference and Where They Matter Inference is fundamentally different from training. Training is compu...

ML Engineering23 min read

Read Chapter →

💡 General

🔥 4× Challenge

179Mesa-Optimization: When Inner Optimizers Emerge

What is MESA and Why Multi-Expert Systems Matter MESA (Modular Expert Scaling Architecture) refers to scaling models usi...

AI Safety23 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

180Deceptive Alignment: The Worst-Case AI Scenario

What is Deceptive Alignment? Deceptive alignment occurs when a model learns an objective different from its training obj...

AI Safety23 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

181AI Governance Frameworks: Regulating Advanced AI

Why AI Governance Matters and What It Covers AI governance encompasses regulations, standards, and self-regulatory pract...

AI Policy23 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

182EU AI Act: Europe's Comprehensive AI Regulation

AI Ethics and Fairness AI systems amplify biases from training data. Algorithmic fairness defines multiple criteria: dem...

AI Policy19 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

183India's AI Regulation Path: Building Frameworks for AI

AI Ethics and Fairness AI systems amplify biases from training data. Algorithmic fairness defines multiple criteria: dem...

AI Policy19 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

184Responsible AI Deployment: From Research to Production

Core Programming Concepts Programming translates human intent into precise machine instructions. Variables store data, f...

AI Systems20 min read

Read Chapter →

🧩 Algorithms

🔥 4× Challenge

185Quantum Computing Basics: Qubits and Quantum Algorithms

Quantum Mechanics Meets Computation: Qubits, Superposition, Entanglement Classical bits are 0 or 1. Quantum bits (qubits...

Quantum23 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

186Quantum Machine Learning: Quantum Advantages in AI

Quantum ML Vs Classical ML: Exponential Advantage or Marketing? Quantum machine learning promises exponential speedups f...

Quantum23 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

187Neuromorphic Computing: Brain-Inspired Architectures

What is Neuromorphic Computing? Neuromorphic computing mimics biological neural systems: spiking neurons, event-driven c...

Hardware23 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

188Analog AI Accelerators: Computing with Physics

Why Analog Computing for AI? Traditional digital computers: discrete 0s and 1s, von Neumann architecture (separate memor...

Hardware22 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

189Uncertainty Quantification in Neural Networks

Confidence and Calibration Neural networks output predictions but don't naturally output confidence. Uncertainty quantif...

ML Theory19 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

190Adversarial Robustness: Defending Against Attacks

Adversarial Examples and Threat Models Small, imperceptible perturbations to inputs can fool neural networks. Add carefu...

ML Security20 min read

Read Chapter →

💡 General

🔥 4× Challenge

191Continual Learning: Learning Without Catastrophic Forgetting

Catastrophic Forgetting Problem When training sequentially on multiple tasks, networks catastrophically forget previous ...

Learning Theory21 min read

Read Chapter →

💡 General

🔥 4× Challenge

192Meta-Learning: Learning How to Learn

Few-Shot Learning and Rapid Adaptation Meta-learning trains models to quickly adapt to new tasks with few examples. Stan...

Learning Theory20 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

193Causal Inference in Machine Learning

Correlation is Not Causation Machine learning finds correlations. "Ice cream sales correlate with drowning deaths" (both...

ML Theory20 min read

Read Chapter →

🤖 AI

🔥 4× Challenge

194Federated Learning: Distributed Privacy-Preserving Training

The Privacy-Centralization Tradeoff Traditional ML: centralize all data on company servers, train models. Privacy risk: ...

Distributed Computing20 min read

Read Chapter →

🤖 AI

Deep Dive

195Building Large Language Models from Scratch: Tokenization to Training

From Text to Numbers Neural networks operate on numbers, not text. "Hello" is meaningless to a neural network. We must c...

AI & Machine Learning22 min read

Read Chapter →

🤖 AI

Deep Dive

196Quantum Computing for AI: The Future of Computation

Beyond Classical Computing Classical computers use bits: 0 or 1. Quantum computers use quantum bits (qubits): can be 0, ...

AI & Machine Learning22 min read

Read Chapter →

🤖 AI

Deep Dive

197Building the Transformer: The Architecture That Changed AI

Building the Transformer: The Architecture That Changed AI Introduction: Why Transformers? Before 2017, the dominant arc...

Advanced Deep Learning27 min read

Read Chapter →

💡 General

Deep Dive

198Scaling Laws: The Mathematical Blueprint Behind GPT-4

Scaling Laws: The Mathematical Blueprint Behind GPT-4 The Fundamental Question When Transformers emerged in 2017, a natu...

Advanced Deep Learning25 min read

Read Chapter →

💡 General

Deep Dive

199RLHF: How ChatGPT Learned to Be Helpful

RLHF: How ChatGPT Learned to Be Helpful The Problem: Next-Token Prediction Isn't Enough GPT-3 was trained with a simple ...

Advanced Deep Learning26 min read

Read Chapter →

🤖 AI

Deep Dive

200Constitutional AI: Making AI Systems Harmless and Honest

Constitutional AI: Making AI Systems Harmless and Honest Beyond RLHF: The Limitations of Feedback RLHF works remarkably ...

Advanced Deep Learning26 min read

Read Chapter →

💡 General

Deep Dive

201Sparse Attention: Making Transformers Efficient at Scale

Sparse Attention: Making Transformers Efficient at Scale The Quadratic Bottleneck: Why Context Length Matters Standard T...

Advanced Deep Learning25 min read

Read Chapter →

🤖 AI

Deep Dive

202Mixture of Experts: Scaling Models Efficiently

Mixture of Experts: Scaling Models Efficiently The Problem: The Compute-Parameter Tradeoff Scaling laws teach us that la...

Advanced Deep Learning27 min read

Read Chapter →

🏆 Grade 12 Quiz — 93 Questions

Test what you've learned with grade-appropriate questions.

Take Grade 12 Quiz →

Other Grades

Grade 8 Grade 9 Grade 10 Grade 11