Modern AI Engineering

AI Engineer Bootcamp · Curriculum

Choose your learning track

Learn machine learning and deep learning, generative AI engineering, or the complete path with career preparation. Every lesson ends with a 5-question quiz.

ML & Deep Learning

How machines learn from data: regression, losses, regularisation, neural networks, backpropagation and the road to Transformers.

3 modules · 20 lessons · 14 interactive labs · from ₹499 per month

Generative AI Engineering

Inside LLMs and the systems around them: Transformers, prompting, RAG, agents, inference, evaluation, safety and infrastructure.

16 modules · 129 lessons · 28 interactive labs · from ₹799 per month

Complete AI Engineer

Everything in both tracks, in order, plus AI engineer career preparation: interview questions, system design and a study plan.

19 modules · 149 lessons · 42 interactive labs · from ₹1,199 per month

Module 1: AI Engineering Starter Kit

Before going deep, we meet the six words that come up in every AI engineering conversation: LLM, RAG, MCP, Agent, Fine-tuning and Quantization.

  1. 1.1 Six Concepts Every AI Engineer Must Know

Module 2: Learning from Data

In this module, we will learn what Machine Learning is, the different ways a machine can learn, and the basic terms we will keep using in every later module of this AI Engineering Course.

  1. 2.1 Machine Learning from First Principles
  2. 2.2 Labeled vs Unlabeled: Two Ways Machines Learn
  3. 2.3 Predicting Numbers vs Categories: Regression Compared
  4. 2.4 Feature Engineering: Turning Raw Data into Signal
  5. 2.5 Precision and Recall: Picking the Right Metric
  6. 2.6 L1 vs L2 Loss: Choosing Your Error Penalty
  7. 2.7 Regularization: Stopping Overfitting with L1 and L2
  8. 2.8 Reinforcement Learning: Teaching Agents Through Reward
  9. 2.9 Contrastive Learning: Training by Comparison

Module 3: Neural Architectures Deep Dive

In this module, we will learn how a neural network actually learns. We will understand the math behind gradient descent and backpropagation step by step, and the techniques that make training stable.

  1. 3.1 Neural Network Bias: What It Is and Why It Matters
  2. 3.2 Gradient Descent: Rolling Downhill to the Optimum
  3. 3.3 Backpropagation: How Neural Networks Learn from Mistakes
  4. 3.4 Cross-Entropy Loss: Scoring Probability Predictions
  5. 3.5 Dropout: Controlled Forgetting as Regularization
  6. 3.6 Batch Norm vs Layer Norm: When to Use Each
  7. 3.7 RMSNorm: Simpler Normalization for Transformers
  8. 3.8 Recurrent Neural Networks: Processing Sequences in Order
  9. 3.9 PyTorch Internals: Dynamic Graphs and Autograd
  10. 3.10 TensorFlow Explained: Static Graphs and Production ML

Module 4: Transformers and How They Think

In this module, we will learn what Generative AI is and how the Transformer, the architecture behind every modern LLM, works from the inside. We will go from tokens to embeddings to attention, one piece at a time.

  1. 4.1 Generative AI: Creating Instead of Classifying
  2. 4.2 Autoregressive Models: Predicting One Token at a Time
  3. 4.3 BPE Tokenization: How LLMs Split Text into Tokens
  4. 4.4 Embeddings: Encoding Meaning as Vectors
  5. 4.5 RNNs vs Transformers: A Fundamental Architecture Shift
  6. 4.6 The Transformer Architecture: Built on Attention
  7. 4.7 Encoder vs Decoder: Two Sides of the Transformer
  8. 4.8 Self-Attention: How Tokens See One Another
  9. 4.9 Attention Math: Queries, Keys, and Values Unpacked
  10. 4.10 Scaled Dot-Product Attention: Why We Divide by √dₖ
  11. 4.11 Causal Masking: Preventing the Model from Seeing the Future
  12. 4.12 Multi-Head Attention: Many Perspectives at Once
  13. 4.13 Cross-Attention: Connecting Encoder Output to the Decoder
  14. 4.14 Rotary Position Encoding: Position Without Fixed Lookup Tables
  15. 4.15 Feed-Forward Networks: The Transformer's Memory Layer

Module 5: Inside the LLM Output Pipeline

In this module, we will learn how an LLM picks the next token, how we control its creativity, how the output reaches the user token by token, and where the context window fails.

  1. 5.1 Temperature Sampling: Dialing Up or Down Creativity
  2. 5.2 Nucleus Sampling: Top-k and Top-p Demystified
  3. 5.3 Token Streaming: Rendering Outputs as They Arrive
  4. 5.4 Lost in the Middle: Why LLMs Miss Central Context

Module 6: Next-Gen LLM Architectures

In this module, we will learn the improvements that modern LLMs add on top of the basic Transformer to become bigger, faster, and able to handle longer inputs. At the end, we will see all of these ideas together inside a real model.

  1. 6.1 A Timeline of LLM Architecture Improvements
  2. 6.2 Mixture of Experts: Routing Tokens to Specialists
  3. 6.3 Grouped Query Attention: Fewer KV Heads, Same Quality
  4. 6.4 Sliding Window Attention: Taming Very Long Contexts
  5. 6.5 Attention Sinks: The Hidden Cost of Extended Context
  6. 6.6 Flash Attention: Memory-Efficient Attention at Scale
  7. 6.7 DeepSeek-V4: Anatomy of an Open-Source Frontier Model

Module 7: The Language Model Zoo

In this module, we will learn that not every language model is a large, text-generating LLM. We will see the smaller, reasoning, recursive, diffusion-based, and decision-only models and when to use which one.

  1. 7.1 Small Language Models: Big Capability in Compact Form
  2. 7.2 Large Reasoning Models: Chain-of-Thought at Inference Time
  3. 7.3 Recursive Language Models: Self-Referential Generation
  4. 7.4 Diffusion Language Models: Text Generation Beyond Autoregression
  5. 7.5 Jev and System One: Fast vs Deliberate AI Thinking

Module 8: Teaching and Shaping Models

In this module, we will learn how a pre-trained model is adapted to our own task, how it is made smaller, and how it is taught to follow instructions and human preferences.

  1. 8.1 Fine-Tuning: Adapting a Pre-Trained Model to Your Task
  2. 8.2 LoRA: Parameter-Efficient Fine-Tuning via Low-Rank Matrices
  3. 8.3 Prefix Tuning: Learnable Context Prepended to the Input
  4. 8.4 Knowledge Distillation: Compressing Large Models into Small Ones
  5. 8.5 Continual Learning: Training Without Forgetting the Past
  6. 8.6 Deep RL from Human Preferences: The Foundational Paper
  7. 8.7 InstructGPT: Teaching GPT-3 to Follow Instructions
  8. 8.8 RLHF: Aligning LLMs with Human Preferences
  9. 8.9 PPO: The Reinforcement Algorithm Behind Instruction Tuning
  10. 8.10 DPO: Alignment Without the Separate Reward Model
  11. 8.11 GRPO: Group-Based Preference Optimization Explained

Module 9: The Art of Prompting

In this module, we will learn how to talk to an LLM so that it gives better answers, and how to manage everything that goes into its context window.

  1. 9.1 Chain-of-Thought Prompting: Making Models Reason Step by Step
  2. 9.2 Prompt Chaining: Decomposing Complex Tasks into Steps
  3. 9.3 Prompt Caching: Reusing Computation Across API Calls
  4. 9.4 Context Engineering: Curating the Model's Working Memory
  5. 9.5 Context Compaction: Fitting More Into a Finite Window

Module 10: Building RAG Systems

In this module, we will learn how to give an LLM knowledge that it was never trained on. We will start with how vectors are stored and searched, then move to retrieval techniques, and finally to the advanced forms of RAG.

  1. 10.1 Vector Databases: Storing and Searching Embeddings at Scale
  2. 10.2 ANN Search: Finding Similar Vectors Without Brute Force
  3. 10.3 Semantic Search: Finding Meaning, Not Just Keywords
  4. 10.4 Hybrid Search: Combining Sparse and Dense Retrieval
  5. 10.5 Rerankers: Re-Scoring Retrieved Results by Relevance
  6. 10.6 ColBERT: Token-Level Late Interaction for Retrieval
  7. 10.7 Document Chunking Strategies for RAG
  8. 10.8 HyDE: Generating Hypothetical Documents to Improve RAG
  9. 10.9 Embedding Caches: Avoiding Redundant Embedding Calls
  10. 10.10 Semantic Caching: Skipping the LLM for Similar Queries
  11. 10.11 Agentic RAG: Dynamic Retrieval with Multi-Step Reasoning
  12. 10.12 GraphRAG: Combining Knowledge Graphs with Retrieval
  13. 10.13 Vectorless RAG: Retrieval Without Embeddings or a Vector Store

Module 11: Autonomous AI Agents

In this module, we will learn how an LLM goes from answering questions to actually doing work. We will start with a single agent, see how it uses tools and memory, and then move to systems where many agents work together.

  1. 11.1 AI Agents: Autonomous Decision-Making Systems
  2. 11.2 Function Calling: Giving LLMs Tools to Act on the World
  3. 11.3 The Agent Loop: Observe, Think, Act, Repeat
  4. 11.4 ReAct Agents: Interleaving Reasoning and Acting
  5. 11.5 Plan-and-Execute: Tackling Complex Tasks in Two Phases
  6. 11.6 Reflection Agents: Self-Critique for Higher-Quality Outputs
  7. 11.7 Agent Memory: Short-Term, Long-Term, and Episodic
  8. 11.8 Model Context Protocol: A Standard Interface for Agent Tools
  9. 11.9 Agent Skills: Reusable Capabilities in Agentic Systems
  10. 11.10 Open Knowledge Format: Structured Agent-to-Agent Communication
  11. 11.11 Multi-Agent Systems: Dividing Work Among Specialist Agents
  12. 11.12 Subagents: Delegating Tasks Within an Agent Network
  13. 11.13 Agent Communication: Protocols and Message Formats
  14. 11.14 AI Orchestration: Coordinating Agents, Tools, and Flows
  15. 11.15 Sakana Fugu: Lessons from an Open-Source Agent Study
  16. 11.16 Computer-Use Agents: Controlling Interfaces with AI

Module 12: Agent Patterns and Frameworks

In this module, we will learn the engineering practices for building reliable agents, and then see how the popular frameworks and coding agents are built.

  1. 12.1 Harness Engineering: The Scaffolding Around AI Agents
  2. 12.2 Loop Engineering: Designing Reliable Agentic Loops
  3. 12.3 Graph Engineering: Stateful Workflows for Agents
  4. 12.4 Defining Done: Why Exit Criteria Shape Agent Quality
  5. 12.5 LangChain: Composable Components for LLM Applications
  6. 12.6 LangGraph: Graph-Based Agent Orchestration Explained
  7. 12.7 Claude Code: AI-Powered Software Engineering at the CLI
  8. 12.8 Cursor: Inside an AI-Native Code Editor

Module 13: Serving LLMs at Scale

In this module, we will learn how to make LLMs faster and cheaper to run. We will start with what happens during inference, then learn the caching, batching, and speculation techniques, then quantization, and finally the serving engines that put it all together.

  1. 13.1 LLM Inference Optimization: The Full Landscape
  2. 13.2 Prefill vs Decode: Two Distinct Phases of LLM Inference
  3. 13.3 Prefill-Decode Disaggregation: Splitting the Two Phases
  4. 13.4 The KV Cache: Avoiding Redundant Attention Computation
  5. 13.5 KV Cache Compression: Trading Some Accuracy for Speed
  6. 13.6 Paged Attention: OS-Inspired Memory Management for KV Caches
  7. 13.7 Continuous Batching: Keeping GPUs Busy Between Requests
  8. 13.8 Speculative Decoding: Draft Fast, Verify in Parallel
  9. 13.9 N-gram Speculation: Draft Tokens Without a Draft Model
  10. 13.10 Medusa: Parallel Decoding via Multiple Prediction Heads
  11. 13.11 EAGLE: Feature-Level Drafting for Faster Inference
  12. 13.12 Model Quantization: Shrinking Weights Without Breaking Outputs
  13. 13.13 GGUF: The File Format Powering Local LLM Inference
  14. 13.14 llama.cpp: Running Large Models on Consumer Hardware
  15. 13.15 vLLM: High-Throughput Serving with PagedAttention
  16. 13.16 SGLang: Structured LLM Programs for Efficient Inference
  17. 13.17 TensorRT-LLM: NVIDIA's Optimized Inference Engine

Module 14: Measuring What Matters

In this module, we will learn how to measure whether our LLM and our agent are actually doing a good job, and how to see what they are doing in production.

  1. 14.1 Evaluating LLMs: Metrics, Benchmarks, and Methods
  2. 14.2 LLM-as-Judge: Automating Evaluation with Another Model
  3. 14.3 Evaluating AI Agents: Metrics and Methods That Work
  4. 14.4 Agent Observability: Traces, Spans, and Debug Signals

Module 15: Securing AI Systems

In this module, we will learn how to keep an LLM application safe, how attackers try to break it, and how AI-generated text can be identified.

  1. 15.1 LLM Guardrails: Filtering Inputs and Outputs for Safety
  2. 15.2 Prompt Injection: Attacks Against LLM-Powered Systems
  3. 15.3 LLM Watermarking: Embedding Invisible Signatures in AI Text

Module 16: Beyond Text: Multimodal AI

In this module, we will learn how AI works with images and other types of data, and the generative models that create images from noise.

  1. 16.1 Multimodal AI: Perceiving Text, Images, and Audio Together
  2. 16.2 Vision Transformers: Applying Self-Attention to Image Patches
  3. 16.3 Image Embeddings: Encoding Visual Content as Vectors
  4. 16.4 Diffusion Models: Iterative Denoising to Generate Images
  5. 16.5 GANs: A Generator and Discriminator in Constant Competition
  6. 16.6 Variational Autoencoders: Learning a Compressed Latent Space

Module 17: Production AI Infrastructure

In this module, we will learn the hardware that runs AI models, where to deploy a model, how to send each request to the right model, and how to design a complete AI system end to end.

  1. 17.1 GPUs for Deep Learning: Parallelism at the Core
  2. 17.2 CUDA Kernels: Writing Parallel Code for NVIDIA GPUs
  3. 17.3 Google TPUs: Purpose-Built Hardware for Neural Networks
  4. 17.4 Language Processing Units: A New Approach to LLM Inference
  5. 17.5 Cloud vs Edge: Where Should Your Model Run?
  6. 17.6 On-Device ML: A TensorFlow Lite Android Walkthrough
  7. 17.7 LLM Routing: Directing Each Query to the Best Model
  8. 17.8 Building a Real-Time Voice AI Agent from Scratch
  9. 17.9 System Design Fundamentals for AI Engineers
  10. 17.10 Transport Protocols: HTTP, WebSockets, and SSE Compared
  11. 17.11 How do Voice And Video Call Work?

Module 18: The Edge of AI Research

In this module, we will learn the ideas that are shaping the future of AI, from models that learn an internal picture of the world to systems that improve themselves.

  1. 18.1 JEPA: LeCun's Vision for World Model AI
  2. 18.2 World Models: Teaching AI to Simulate Its Environment
  3. 18.3 Recursive Self-Improvement: Can AI Improve Itself Indefinitely?

Module 19: AI Engineering Career Prep

We have learned everything from machine-learning foundations to AI agents in production. Now we turn that knowledge into clear interview answers and confident system designs.

  1. 19.1 Cracking the AI Engineering Interview