Executive Post Graduate Certification in Data Science & Artificial Intelligence

Python 

• Introduction to Python and IDEs – The basics of the python programming language, how you can use various IDEs for python development like Jupyter, Pycharm, etc. 
• Python Basics – Variables, Data Types, Loops, Conditional Statements, functions, decorators, lambda functions, etc.
• Object Oriented Programming – Introduction to OOPs concepts like classes, objects, inheritance, abstraction, polymorphism, encapsulation, etc.
• Hands-on Sessions And Assignments for Practice – The culmination of all the above concepts with real-world problem statements for better understanding. 

Linux

• Introduction to Linux  – Establishing the fundamental knowledge of how linux works and how you can begin with Linux OS. 
• Linux Basics – File Handling, data extraction, etc.
• Hands-on Sessions And Assignments for Practice – Strategically curated problem statements for you to start with Linux. 

SQL Basics – 

• Fundamentals of Structured Query Language
• SQL Tables, Joins, Variables 

Advanced SQL –  

• SQL Functions, Subqueries, Rules, Views
• Nested Queries, string functions, pattern matching
• Mathematical functions, Date-time functions, etc. 

Deep Dive into User Defined Functions

• Types of UDFs, Inline table value, multi-statement table. 
• Stored procedures, rank function, SQL ROLLUP, etc.

SQL Optimization and Performance

• Record grouping, searching, sorting, etc. 
• Clustered indexes, common table expressions.

Hands-on exercise: 

Writing comparison data between past year to present year with respect to top products, ignoring the redundant/junk data, identifying the meaningful data,  and identifying the demand in the future(using complex subqueries, functions, pattern matching concepts).

Statistics and Descriptive Analytics using MS Excel

• Measure of central tendency, measure of spread, five points summary, etc.
• Probability Distributions, Probability in Business Analytics.
• Probability Distributions, Binomial distribution, Poisson distribution, bayes theorem, central limit theorem.

Python for Descriptive, Diagnostic, and Inferential Statistics

• Correlation, covariance, confidence intervals, hypothesis testing, F-test, Z-test, t-test, ANOVA, chi-square test, etc.

Prescriptive Analytics

Introduction to Machine learning 

• Supervised, Unsupervised learning.
• Introduction to scikit-learn, etc.

Regression 

• Introduction classification problems, Identification of a regression problem, dependent and independent variables.
• How to train the model in a regression problem.
• How to evaluate the model for a regression problem.
• How to optimize the efficiency of the regression model.

Classification 

• Introduction to classification problems, Identification of a classification problem, dependent and independent variables.
• How to train the model in a classification problem.
• How to evaluate the model for a classification problem.
• How to optimize the efficiency of the classification model.

Clustering 

• Introduction to clustering problems, Identification of a clustering problem, dependent and independent variables.
• How to train the model in a clustering problem.
• How to evaluate the model for a clustering problem.
• How to optimize the efficiency of the clustering model.

• Linear Regression – Creating linear regression models for linear data using statistical tests, data preprocessing, standardization, normalization, etc.
• Logistic Regression – Creating logistic regression models for classification problems – such as if a person is diabetic or not, if there will be rain or not, etc.
• Decision Tree – Creating decision tree models on classification problems in a tree like format with optimal solutions.
• Random Forest – Creating random forest models for classification problems in a supervised learning approach.
• Support Vector Machine – SVM or support vector machines for regression and classification problems.
• K-Nearest Neighbors – A simple algorithm that can be used for classification problems.
• Time Series Forecasting – Making use of time series data, gathering insights and useful forecasting solutions using time series forecasting.

Performance Metrics

• Classification reports – To evaluate the model on various metrics like recall, precision, f-support, etc.
• Confusion matrix – To evaluate the true positive/negative, false positive/negative outcomes in the model.
• r2, adjusted r2, mean squared error, etc.

• K-means – The k-means algorithm that can be used for clustering problems in an unsupervised learning approach.
• Dimensionality reduction – Handling multi dimensional data and standardizing the features for easier computation.
• Linear Discriminant Analysis –  LDA or linear discriminant analysis to reduce or optimize the dimensions in the multidimensional data.
• Principal Component Analysis – PCA follows the same approach in handling the multidimensional data.

Bagging And Boosting Algorithms

• XGBoost – Extreme Gradient Boosting or XGBoost algorithm works on an ensemble approach that combines the predictions of weak models to produce a strong prediction.
• ADAboost – An ensemble machine learning algorithm that converts the weak learners to strong learners for better performance.
• Gradient Tree Boosting – One of the powerful boosting algorithms works on the principle of reducing the bias in predictions.
• Bootstrap Aggregation – A bagging technique that fits decision trees on bootstrap samples of the training data.

Other Machine Learning Algorithms

• Ordinary Least Square – A simple regression algorithm that is used to estimate the coefficient of the linear regression equation for predictions.
• Markov Chain – A stochastic approach or algorithm that describes the possible chain of events in terms of probability theory.
• Naive Bayes – Bayes theorem inspired algorithm in machine learning that is used for predictive modeling.
• Stochastic Gradient Descent – A variant of Gradient Descent that is used to optimize machine learning algorithms.
• Gaussian Mixture Model – A probabilistic model that assumes the data points are generated from a mix of gaussian distributions with unknown parameters.
• Singular Value Decomposition – Algorithm based on factorization of a real or complex matrix, that is often used to create a recommendation engine.

Predictive Analytics And Machine Learning

• Regression and Multivariate Analysis 
• Classification problems in machine learning 
• Data Multidimensionality and Linear Algebra
• Feature engineering and Feature selection 
• Hyperparameter Tuning and other optimization techniques

Cognitive Science and Analytics

• Understanding Natural Language Processing Applications (e.g. Search Engines and Social Media)
• Machine Learning Applications and Chatbots
• Social Media Analytics Advanced Text Mining like Sentiment Analysis, Topic Modelling, and Text Summarisation

Artificial Intelligence Basics 

• Introduction to keras API and tensorflow

Neural Networks

• Single Cell (perceptron)
• Multi cell perceptron Topology
• Weights & Biases
• Build a NN from scratch (using numpy)

Deep Learning 

• Use cases of DL in industry
• Difference between DS, ML, DL & AI
• Lifecycle of Deep Learning Project

RBM and DBNs & Variational AutoEncoder

• Introduction rbm and autoencoders
• Deploying rbm for deep neural networks, using rbm for collaborative filtering
• Autoencoders features and applications of autoencoders.

Object Detection using Convolutional Neural Net

• Constructing a convolutional neural network using TensorFlow
• Convolutional, dense, and pooling layers of CNNs
• Filtering images based on user queries

Generating images with Neural Style and Working with Deep Generative Models

• Automated conversation bots leveraging
• Generative model, and the sequence to sequence model (lstm).

Distributed & Parallel Computing for Deep Learning Models

• Parallel Training, Distributed vs Parallel Computing
• Distributed computing in Tensorflow, Introduction to tf.distribute
• Distributed training across multiple CPUs, Distributed Training
• Distributed training across multiple GPUs, Federated Learning
• Parallel computing in Tensorflow

Reinforcement Learning

• Mapping the human mind with deep neural networks (dnns)
• Several building blocks of artificial neural networks (anns)
• The architecture of dnn and its building blocks
• Reinforcement learning in dnn concepts, various parameters, layers, and optimization algorithms in dnn, and activation functions.

Deploying Deep Learning Models and Beyond

• Understanding model Persistence, Saving and Serializing Models in Keras, Restoring and loading saved models
• Introduction to Tensorflow Serving, Tensorflow Serving Rest, Deploying deep learning models with Docker & Kubernetes, Tensorflow Serving Docker, Tensorflow Deployment Flask.
• Deploying deep learning models in Serverless Environments
• Deploying Model to Sage Maker
• Explain Tensorflow Lite Train and deploy a CNN model with TensorFlow

Version Control 

• What is version control, types, SVN.

GIT

• Git Lifecycle, Common Git commands, Working with branches in Git
• Github collaboration (pull request), Github Authentication (ssh and Http)
• Merging branches, Resolving merge conflicts, Git workflow

1. Non-Relational Data Stores and Azure Data Lake Storage

1.1 Document data stores
1.2 Columnar data stores
1.3 Key/value data stores
1.4 Graph data stores
1.5 Time series data stores
1.6 Object data stores
1.7 External index
1.8 Why NoSQL or Non-Relational DB?
1.9 When to Choose NoSQL or Non-Relational DB?
1.10 Azure Data Lake Storage

Definition, Azure Data Lake-Key Components, How it stores data? Azure Data Lake Storage Gen2, Why Data Lake? Data Lake Architecture

2. Data Lake and Azure Cosmos DB

2.1 Data Lake Key Concepts
2.2 Azure Cosmos DB
2.3 Why Azure Cosmos DB?
2.4 Azure Blob Storage
2.5 Why Azure Blob Storage?
2.6 Data Partitioning: Horizontal partitioning, vertical partitioning, Functional partitioning
2.7 Why Partitioning Data?
2.8 Consistency Levels in AzureCosmos DB:  Semantics of the five-consistency level

3. Relational Data Stores

3.1 Introduction to Relational Data Stores
3.2 Azure SQL Database – Deployment Models, Service Tiers
3.3 Why SQL Database Elastic Pool?

4. Why Azure SQL?

4.1 Azure SQL Security Capabilities
4.2 High-Availability and Azure SQL Database: Standard Availability Model, Premium Availability Model
4.3 Azure Database for MySQL
4.4 Azure Database for PostgreSQL
4.5 Azure Database for MariaDB
4.6 What is PolyBase and Why PolyBase?
4.7 What is Azure Synapse Analytics (formerly SQL DW): SQL Analytics and SQL pool in Azure Synapse, Key component of a big data solution, SQL Analytics MPP architecture components

5. Azure Batch

5.1 What is Azure Batch?
5.2 Intrinsically Parallel Workloads
5.3 Tightly Coupled Workloads
5.4 Additional Batch Capabilities
5.5 Working of Azure Batch

6. Azure Data Factory

6.1 Flow Process of Data Factory
6.2 Why Azure Data Factory
6.3 Integration Runtime in Azure Data Factory
6.4 Mapping Data Flows

7. Azure Data Bricks

7.1 What is Azure Databricks?
7.2 Azure Spark-based Analytics Platform
7.3 Apache Spark in Azure Databricks

8. Azure Stream Analytics

8.1 Working of Stream Analytics
8.2 Key capabilities and benefits
8.3 Stream Analytics Windowing Functions: Tumbling window, Hopping Window, Sliding Window, Session Window

Introduction to MLOps 

• MLOps lifecycle
• MLOps pipeline 
• MLOps Components, Processes, etc

Deploying Machine Learning Models 

• Introduction to Azure Machine Learning 
• Deploying Machine Learning Models using Azure

The Data Science capstone project focuses on establishing a strong hold of analyzing a problem and coming up with solutions based on insights from the data analysis perspective. The capstone project will help you master the following verticals: 

• Extracting, loading and transforming data into usable format to gather insights. 
• Data manipulation and handling to pre-process the data.
• Feature engineering and scaling the data for various problem statements. 
• Model selection and model building on various classification, regression problems using supervised/unsupervised machine learning algorithms.
• Assessment and monitoring of the model created using the machine learning models.

• Recommendation Engine – The case study will guide you through various processes and techniques in machine learning to build a recommendation engine that can be used for movie recommendations, restaurant recommendations, book recommendations, etc.
• Rating Predictions – This text classification and sentiment analysis case study will guide you towards working with text data and building efficient machine learning models that can predict ratings, sentiments, etc.
• Census – Using predictive modeling techniques on the census data, you will be able to create actionable insights for a given population and create machine learning models that will predict or classify various features like total population, user income, etc.
• Housing – This real estate case study will guide you towards real world problems, where a culmination of multiple features will guide you towards creating a predictive model to predict housing prices.
• Object Detection – A much more advanced yet simple case study that will guide you towards making a machine learning model that can detect objects in real time.
• Stock Market Analysis – Using historical stock market data, you will learn about how feature engineering and feature selection can provide you some really helpful and actionable insights for specific stocks.
• Banking Problem – A classification problem that predicts consumer behavior based on various features using machine learning models.
• AI Chatbot – Using the NLTK python library, you will be able to apply machine learning algorithms and create an AI chatbot.