Founder & CEO @ Writesonic, Botsonic and Socialsonic | Forbes30U30 | AI Content Marketing and Chatbots

Samanyou Garg  • Peerlist

• Implemented a Naive Bayes Classifier to distinguish between 'spam', 'non-spam' and 'uncertain' email messages. Also, analysed the performance on 3 different training/testing data sets and optimised the model to achieve better classification results. 
• Created a Nearest Centroid Classifier (also known as Rocchio Classifier) and tested its performance with the same 3 data sets.
• Analysed the performance of the Random Forest (used Matlab's TreeBagger class) for spam filtering problem using different number of decision trees.

Machine Learning - Spam Filtering Project

EventLite - Event Management Made Simple

• Developed software from first principles to create a new web application using Spring web framework as a team of 5 people over 10 weeks.
• Introduced to MVC system architecture with Spring.
• Started by creating User Interface Mockups for the initial requirements given.
• Designed for Testability. Constructed good automated tests from requirements, looking in particular at the importance of working out how to ensure coverage, ensuring that functionality can be tested in isolation.
• Integrated Third-Party APIs like Twitter and Google Maps API to use their services in our application.
• Ensured that acceptance tests are met, and the user experience is good.
• Implemented a REST API to enable sharing of data between applications.

Software Engineering - Semester 2 Team Project

Stendhal Game stendhalgame.org
Stendhal is a multi-player online adventure game with over ten years of commit history. 

•	Tested, debugged, built, developed and deployed a multi-user, multi-threaded, client-server open source game containing over 6000 Java classes and 1800 test cases using Eclipse, Git, Apache Ant, Jenkins, Junit and SonarQube. Worked on the automation of builds and tests leading a team of 6 people over 12 weeks. 

Software Engineering - Semester 1 Team Project

• Used 500 digital samples of each digit provided by university to examine feature extraction techniques in image processing and 3 different classifiers, as well as their performance and classification accuracies. 
• Used Matlab language and environment to analyse the performance and testing accuracy of the k-Nearest Neighbour Classifier in a handwritten digit recognition problem. 
• Researched and implemented diagonal-based feature extraction technique to decrease a number of original features from 256 (16x16 pixel images) to 24, and used KNN Classifier to test it 
• Implemented the Perceptron Algorithm to classify two handwritten digits 3 and 
8 (in a digital format)
• Created a Naive Bayes Classifier to classify same 2 handwritten digits

Machine Learning - Handwritten Digit Recognition Project

Bhagavad Gita is a 700 verse Hindu scripture in Sanskrit that is part of the Hindu epic Mahabharata (chapters 23–40 of the 6th book of Mahabharata). It is a practical guide to one's life that guides one to re-organize their life, achieve inner peace and approach the Supreme Lord (the Ultimate Reality).

BhagavadGita.io is a modern Gita app with a simple, beautiful and easy to use interface, helping the user focus on reading.

It takes design inspiration from Google’s material design, which gives it a modern look and feel and also helps it provide a more engaging user experience with perfect usability. Material Design’s reusable elements help it cut page load times without compromising on the design which means a quicker, leaner platform for both the readers and the makers.

BhagavadGita.io perfectly combines this intuitive, user-friendly design with best-in-class features to significantly enhance the experience of its users.

• Implemented OAuth 2.0 authentication to make the app and the API secure.
• Implemented 3rd party login APIs allowing users to sign-in using their social accounts.
• Used Elasticsearch for full-text search
• Implemented Progress Tracking feature, allowing users to keep track of their Chapter-wise, Overall and Monthly Progress using detailed graphs and statistics.
• Implemented Achievements system to reward the users based on their progress.
• Implemented User customizable Reading Plans, allowing them to quickly read Gita at your their pace and improve their Gita reading habits.
• Implemented Favourites feature, allowing users to share, memorize and find their favorite verses.

BhagavadGita.io (Bhagavad Gita --> Simplified)

Budge It: Student budgeting made simple
As a university student, staying on top of finances can be tricky. Budge It is an invaluable tool to help students prioritize their spending and manage their money—no matter how much or how little they have. 
It will help them identify wasteful expenditures, adapt quickly as their financial situation changes, and achieve their financial goals.   

• It is an online budget management web application to help university students keep track of their incomes/expenses, and provided functionality for its analysis in a really simplified form. 
• The role involved project management, UI design and work on the front end.
• The project was conducive towards enhancing my communication, teamwork and problem solving skills.

Unique Selling Points 
• Cashflow Forecast – Uses a simple algorithm to populate a graph and predict when you may run out of cash based on your monthly incomes/expenses. This allows you to identify potential shortfalls in cash balance in advance.
• Secure Login - SHA-256 hashing for secure password storage.