Each module is self-contained, but the recommended sequence ensures every concept has its prerequisites covered. Students targeting specific roles can follow the career-specific paths in the Career Paths section below.
Python, SQL, stats, EDA, ML basics
Algorithm depth, XGBoost, feature eng.
PyTorch, CNNs, Transformers, LoRA
Text processing, BERT, RAG, agents
OpenCV, YOLO, ViT, SAM, diffusion
LLM APIs, RAG, agents, multimodal
MLflow, Docker, CI/CD, monitoring
Each module card shows: what you'll learn, key technologies, duration, and which career paths it feeds. Click "Explore Syllabus" for the full topic-by-topic breakdown.
The foundation module — Python data stack (NumPy, Pandas), SQL for analytics, statistics and probability, exploratory data analysis, supervised and unsupervised ML with scikit-learn, XGBoost, model evaluation, GenAI integration (LLM APIs, RAG basics), and model deployment with Streamlit and FastAPI. The complete skill set for Data Analyst and entry-level Data Scientist roles.
Deep dive into ML algorithms, their mathematics, and engineering. Linear algebra and calculus connected to real algorithms, feature engineering and scikit-learn pipelines, all supervised and unsupervised algorithms with internals (not just API calls), XGBoost/LightGBM/CatBoost mastery, hyperparameter tuning with Optuna, SHAP explainability, NLP with Transformers, responsible AI, and MLOps deployment.
Neural networks from scratch in PyTorch — training loops, CNNs for vision, RNNs/LSTMs for sequences, Transformers (BERT, GPT) from the architecture up. Vision Transformers (ViT), LoRA/QLoRA fine-tuning, GANs, diffusion models (Stable Diffusion), multimodal models (CLIP, LLaVA), RAG architecture, self-supervised learning, and production deployment with quantisation and edge overview.
Text processing to LLM-powered applications — structured across three eras: Classical NLP (tokenisation, TF-IDF, spaCy, scikit-learn classification), Transformer Era (BERT, GPT, Hugging Face fine-tuning, QA, NER), and LLM-Powered NLP (prompt engineering, RAG with evaluation, conversational AI with guardrails, LLM agents, LoRA fine-tuning, Whisper speech integration).
OpenCV image processing, CNN architectures (ResNet → EfficientNet), Vision Transformers (ViT + LoRA), YOLO object detection (v5 → v11), image segmentation (U-Net, SAM), generative vision (GANs, Stable Diffusion), multimodal (CLIP, LLaVA, Florence-2), video tracking, OCR/document AI, Grad-CAM interpretability, and edge deployment (ONNX, TensorRT).
The complete GenAI engineering stack — LLM API integration (OpenAI, Anthropic, Google), RAG pipelines (LangChain, ChromaDB, hybrid search, evaluation), AI agents (LangGraph, CrewAI, multi-agent systems), LLM fine-tuning (LoRA/QLoRA), multimodal AI (CLIP, Stable Diffusion), audio (Whisper, ElevenLabs), MLOps for GenAI, and enterprise application architecture.
Taking models from notebook to production — experiment tracking (MLflow, W&B), model serving (FastAPI, TorchServe, Streamlit, Gradio), Docker containerisation, CI/CD for ML (GitHub Actions), cloud deployment (AWS SageMaker, Bedrock, Azure AI), model monitoring and drift detection (Evidently), cost optimisation, and the LLMOps extension (prompt versioning, evaluation pipelines, guardrails).
Analyse data, create dashboards, tell stories with numbers. SQL + Pandas + visualisation + statistics. The highest-volume entry-level data role. Modules: Data Science
Build predictive models, run experiments, drive business decisions with ML. The role that combines statistics, programming, and domain expertise.Modules: Data Science + Machine Learning
Deploy and scale ML models in production. Build ML pipelines, monitor drift, optimise inference. The bridge between data science and engineering.Modules: Data Science + ML + MLOps
Build LLM-powered applications: RAG systems, AI agents, multimodal apps. The fastest-growing and highest-paying AI role in 2025–26.Modules: Data Science + Deep Learning + GenAI
Build vision systems for autonomous driving, medical imaging, manufacturing, retail. Detection, segmentation, and multimodal understanding.Modules: Data Science + DL + Computer Vision
Build text processing systems, chatbots, document analysis, and language AI. Fine-tune BERT, build RAG, deploy conversational AI.Modules: Data Science + DL + NLP
India needs 1M+ data/AI professionals by 2026 — current supply is under 300K. Every industry (banking, healthcare, e-commerce, manufacturing, telecom) is hiring. The supply-demand gap means qualified candidates get multiple offers at premium packages.
Data Scientist/ML Engineer/AI Engineer roles offer ₹8–40L+ starting — 2–5x higher than generic developer roles at the same experience level. The specialisation premium is massive and growing as AI adoption accelerates across industries.
AI Engineer / GenAI Developer didn't exist as a job title 2 years ago. Now it's the fastest-growing technical role globally. Students who learn RAG, agents, and LLM integration NOW are positioned for a career path that will define the next decade of tech.
Data science isn't limited to tech companies. Banking (risk, fraud), healthcare (diagnostics), agriculture (yield prediction), retail (recommendations), government (policy analysis) — every sector needs data professionals. Geographic mobility is unmatched.
E-305, Industrial Area, Sahibzada Ajit Singh Nagar, Punjab 160055, IN
