Data Science with Agentic AI.
Model. Automate. Evaluate.

Python is the language that we use for scripting and learning Data Science algorithms throughout the course. This section provides a introduction to the basics of Python programming language. Python Basics includes Python Syntax, Output, Comments, Variables, Data Types, Numbers, Casting, Strings, Booleans, Operators, If...Else, Match, While Loops, For Loops, Range, and String Formatting. These topics help learners understand the foundation of Python programming.

Functions are the programming paradigm that increases code reusability in our scripts. This section introduces functions in Python along with other high level concepts like Generators and Decorators.

Data Structures are the skeleton of computer programming. Python In-Built Data Structures include Python Lists, Tuples, Sets, Dictionaries, and Arrays. These topics focus on storing, managing, and manipulating data in different formats.

OOP includes Python Classes, OOP concepts, Classes/Objects, init Method, self Parameter, Class Properties, Class Methods, Inheritance, Polymorphism, Encapsulation, and Inner Classes. These topics teach how to build programs using objects and real-world models.

In this section, we look into file handling with python. We also learn how to safely handle errors in our python programs.

In this section, we look back to functions and learn Generators in detail. We see the iteration protocol in python and implement different iterators in python.

Basics of Machine Learning includes Introduction to Machine Learning, Roadmap to Learning Machine Learning, and Types of Data and Variables in ML. It also covers Data Cleaning techniques such as identifying and handling missing values, One Hot Encoding, Label Encoding, Ordinal Encoding, Outlier Detection and Removal, Feature Scaling using Standardization and Normalization, and handling duplicate data. These topics help learners prepare clean, structured, and high-quality data for building accurate machine learning models.

This section introduces us to Optimisation methods. Also this also introduces some of the basic loss function that we will use and their derivation.

Linear Regression is one of the widely used algorithm in data analytics and data science firms. In this section we learn Linear Regression and its in-depth analogies.

Learn how to map out any straight line i.e. linear in nature with any slope and constant using Linear Regression.

Linear Regression is one of the widely used algorithm in data analytics and data science firms. In this section we learn Linear Regression and its in-depth analogies.

Learn to predict the pricing of any house in your neighbourhood according to realistic factors like size of land, vintage, industry etc. using Multivariate Linear Regression.

How about a Linear model for classification tasks? This section, we learn Logistic Regression and see how classification is performed in Machine Learning.

Learn how to predict the chances of survival of any passenger present in the Titanic ship that is about to sink using all the features and the meta data.

Supervised Machine Learning includes Train-Test Split in datasets, Regression Analysis such as Simple Linear Regression, Multiple Linear Regression, and Polynomial Regression, along with Cost Function, R Squared Score, and Adjusted R Squared for performance evaluation. It also covers Classification techniques including Logistic Regression for Binary and Multiclass problems, Confusion Matrix, handling imbalanced datasets, and the Naive Bayes algorithm. In addition, it includes Non-Linear Supervised Algorithms such as Decision Tree for Classification and Regression, K-Nearest Neighbors for Classification, Hyperparameter Tuning, and Cross-Validation for improving model accuracy and reliability.

Unsupervised Machine Learning includes concepts of Unsupervised Learning and Clustering techniques such as K-means Clustering, Hierarchical Clustering, and DBSCAN Clustering Algorithm. It also covers the Silhouette Score for evaluating clustering performance and measuring the quality of clusters. These methods help in discovering hidden patterns and structures in unlabeled data.

Ensemble Learning in Machine Learning includes the concept of Ensemble Learning and techniques such as Max Voting, Averaging, and Weighted Average Voting for combining multiple models. It also covers Bagging methods using the Bagging Meta-Estimator and Random Forest algorithm to improve prediction accuracy, reduce overfitting, and build more reliable machine learning models.

Deep Learning covers Perceptrons including Single Layer and Multilayer Perceptrons, loss functions such as Hinge Loss and Binary Cross Entropy. It includes Artificial Neural Networks with training techniques like Vanishing and Exploding Gradients, Dropout, Early Stopping, and L1/L2 Regularization, Normalization, and Optimizers such as Exponential Moving Average, SGD with Momentum, AdaGrad, and Adam. It explains Convolutional Neural Networks including convolution, padding, strides, backpropagation, and transfer learning. It also includes Recurrent Neural Networks, their types and challenges, LSTM and Transformers and Attention Mechanisms including encoder-decoder architecture, positional encoding.

SQL covers Introduction to SQL, SQL Data Types, Database and Table Management, and Data Manipulation Language for working with data. It includes SQL Clauses for query refinement, SQL Operators, SQL Functions, and SQL Joins and Subqueries for combining and analyzing data. This section also covers SQL Window Functions along with practical problem-solving.

Mathematics for Data Science covers Introduction to Statistics including Population and Sample concepts, Descriptive and Inferential Statistics, and Basic Statistical Measures. It includes Measures of Central Tendency such as Mean, Median, and Mode, Measures of Variability, Percentages, Percentiles, and Quartiles. This section also covers Probability, Probability Distributions, Normal Distribution, Covariance, and Correlation. Additionally, it includes the Central Limit Theorem and Hypothesis Testing methods such as Confidence Intervals, t-test, Chi-square test, and ANOVA test.

Generative AI covers concepts such as LangChain and Retrieval-Augmented Generation (RAG) pipelines. This section focuses on building intelligent applications by integrating large language models with external data sources, enabling context-aware responses, document retrieval, and end-to-end AI-powered workflows.

NLP covers text preprocessing and normalization, regular expressions (Regex), tokenization, stemming and lemmatization, stop-word removal, Bag-of-Words (BoW), TF-IDF, N-Grams, word embeddings (Word2Vec, GloVe, FastText), sentiment analysis, intent classification, sequence modeling using RNNs, LSTMs, and GRUs.

LLMs covers Transformer architecture, Self-Attention, Multi-Head Attention, Positional Encoding, BERT and GPT architectures, tokenization, pretraining vs. fine-tuning, basic LoRA, and running local LLMs using Ollama.

Covers Generative AI fundamentals, foundation models, prompt engineering (Zero-Shot, Few-Shot), system prompts, temperature and Top-P, structured outputs (JSON, Pydantic), function/tool calling, multimodal inputs, and API integration with LLMs.

Covers Text embeddings, cosine similarity, vector databases (ChromaDB), semantic search, metadata filtering, hybrid search (BM25 + vectors), document parsing, text chunking, RAG pipelines, reranking, source citations, and a Document Question-Answering Chatbot project.

Agentic AI foundation covers AI agents, perception, memory, planning, task decomposition, ReAct (Reasoning and Acting), custom tools using Python functions and APIs, self-correction, error handling, retries, and Human-in-the-Loop (HITL) approval.

LangChain and LangGraph fundamentals, stateful workflows, StateGraph, conditional routing, checkpoints and persistence, basic multi-agent collaboration, MCP (Model Context Protocol) fundamentals, prompt injection defense, and basic agent evaluation using Ragas or DeepEval.

Build a retail analytics system using PostgreSQL or MySQL to analyze customer lifetime value (CLV), purchase patterns, sales trends, and monthly customer retention cohorts. Apply database schema design, advanced SQL queries, CTEs, and window functions to generate business insights.

Analyze more than a century of Olympic records to identify top-performing countries, medal trends, athlete achievements, and performance patterns. Use SQL aggregations, recursive CTEs, ranking functions such as DENSE_RANK, and LAG/LEAD to perform historical data analysis.

Build a modular Python ETL application that extracts data from CSV and JSON files, validates schemas, handles missing or malformed records, transforms data, and exports cleaned datasets with audit logs. Apply OOP, Pydantic validation, logging, command-line interfaces, and unit testing with pytest.

Develop a machine learning model to identify fraudulent credit card transactions in highly imbalanced datasets. Perform exploratory data analysis, feature engineering, class imbalance handling, and model evaluation using precision, recall, F1-score, and Precision-Recall AUC. Use scikit-learn, pandas, NumPy, and imbalanced-learn.

Build a multi-class classification model that predicts possible disease categories from structured symptom data. Perform data preprocessing, feature selection, decision tree classification, ensemble modeling, and cost-complexity pruning. Evaluate performance using confusion matrices, classification reports, and feature importance analysis.

Develop a deep learning model to classify chest X-ray images for pneumonia detection. Apply image preprocessing, data augmentation, convolutional neural networks, transfer learning, and model evaluation with a focus on sensitivity and false-negative reduction. Use PyTorch or TensorFlow/Keras, torchvision, and OpenCV.

Use the sensor data from accelerometers and gyroscopes to classify activities such as walking, resting, and climbing stairs. Learn time-series preprocessing, sliding-window techniques, LSTM models, and performance evaluation.

Build a document-based AI Study Assistant using Python, an LLM API, and LangChain. The application should accept study notes or a document, answer questions using the provided content through a basic RAG pipeline, generate a short quiz, and provide relevant source references. Test the application for answer accuracy and handle questions that cannot be answered from the document.

Build a chatbot that answers questions from multiple PDF documents. Learn document extraction, text chunking, embeddings, vector databases, semantic search, and RAG pipelines to generate answers based on the uploaded documents, with relevant source citations.

Building an AI assistant integrating RAG, tool calling, LangGraph workflows, MCP-based tool integration, conversation memory, safety guardrails, and FastAPI deployment.