The full 78-week curriculum

1. Phase 1 · Mathematical & Programming Foundations

   20 weeks

   Algebra, calculus, vectors, matrices, probability, and Python — the foundations everything else stands on.

   • W1Algebra Refresher + Python Setup
   • W2Functions, Logs, Trig + Python Data Structures
   • W3Limits & Derivatives + Python Functions
   • W4Chain Rule + Object-Oriented Programming
   • W5Integration + Modules, Errors, Environments
   • W6Multivariable Calculus + Files, Testing
   • W7Linear Algebra I — Vectors + Git Deeply
   • W8Linear Algebra II — Matrices + NumPy Intro
   • W9Linear Algebra III — Spaces & Eigen + NumPy Mastery
   • W10SVD & Decompositions + Pandas Intro
   • W11Probability I + Matplotlib
   • W12Probability II + EDA Workflow
   • W13Probability III + Statistical Functions
   • W14Statistics + SQL Basics
   • W15Information Theory + Advanced SQL
   • W16Optimization Basics + SWE Hygiene
   • W17Constrained Optimization + Docker
   • W18Math Consolidation Week
   • W19Programming Capstone
   • W20Phase 1 Buffer + Honest Self-Assessment

2. Phase 2 · Classical Machine Learning

   12 weeks

   Linear & logistic models, trees, ensembles, SVMs, clustering, and honest evaluation on real datasets.

   • W1ML Foundations
   • W2Linear Regression
   • W3Logistic Regression & Classification
   • W4Evaluation & Model Selection
   • W5Decision Trees
   • W6Ensembles I — Bagging & Random Forests
   • W7Ensembles II — Boosting
   • W8SVMs & Kernels
   • W9Clustering
   • W10Dimensionality Reduction
   • W11Naive Bayes, k-NN, Anomaly Detection
   • W12Feature Engineering + Phase 2 Capstone

3. Phase 3 · Deep Learning

   16 weeks

   Neural networks, backprop, CNNs, RNNs, and transformers — built up from a single neuron.

   • W1Neural Network Foundations
   • W2Backpropagation Deeply
   • W3PyTorch Fundamentals
   • W4Training Dynamics
   • W5Optimization for Deep Learning
   • W6CNNs I
   • W7CNN Architectures
   • W8Sequence Models (RNNs & LSTMs)
   • W9Attention & Transformers I
   • W10Transformers II — Building GPT
   • W11Tokenization Deeply
   • W12Pretraining Concepts
   • W13Fine-tuning & Transfer Learning
   • W14Parameter-Efficient Fine-Tuning (PEFT)
   • W15Computer Vision Deep Dive (Optional Specialization)
   • W16Phase 3 Capstone

4. Phase 4 · Generative AI

   14 weeks

   LLMs, embeddings, RAG, diffusion, and multimodal models — how generative AI actually works.

   • W1Generative Models Overview + Autoencoders
   • W2Variational Autoencoders
   • W3GANs
   • W4Diffusion I — Foundations
   • W5Diffusion II — Modern Variants
   • W6LLM Internals
   • W7RLHF & Alignment
   • W8Prompt Engineering & Reasoning
   • W9Embeddings & Vector Stores
   • W10Retrieval-Augmented Generation (RAG)
   • W11RAG Evaluation
   • W12Multimodal Models
   • W13Inference Optimization
   • W14GenAI Capstone

5. Phase 5 · Agentic AI

   10 weeks

   Tools, planning, memory, MCP, and multi-agent systems — turning models into agents.

   • W1Agent Foundations
   • W2Tool Use & Function Calling
   • W3Planning & Reasoning
   • W4Memory Systems
   • W5Agent Frameworks I — LangGraph
   • W6Agent Frameworks II — Survey
   • W7MCP (Model Context Protocol)
   • W8Multi-Agent Systems
   • W9Agent Evaluation & Observability
   • W10Agent Safety

6. Phase 6 · Production, Research & Capstone

   6 weeks

   MLOps, serving, monitoring, and a real build-and-ship capstone that becomes your portfolio.

   • W1MLOps Foundations
   • W2Model Serving & Deployment
   • W3Monitoring & Drift
   • W4Capstone I — Scope, Plan & Data
   • W5Capstone II — Build the End-to-End System
   • W6Capstone III — Ship, Document & Present

Three depth-tuned tracks (Class 9-10, 11-12, Graduation) — same curriculum, depth adjusted. English throughout; Hindi & Hinglish on selected early weeks.