Arshdeep Singh Sachdeva's Projects
This repository is being used to personalize my github profile.
Personal Portfolio Website demonstrating my background, projects and my professional expereience.
A react application to upload, record audio files and then mix and merge them to create a new file.
This project is in-progress, I am using AWS tools to build an image/video detection and processing application
This project focuses on developing a credit card fraud detection system using logistic regression. The process involved preprocessing the raw dataset, handling missing values, addressing outliers, and normalizing features, traning and testing the model.
This data visualization project underscores my proficiency in both data visualization techniques and frontend development. Leveraging a skill set that includes D3, Javascript, Bootstrap, CSS, and HTML, I have successfully crafted an interactive and dynamic visualization of the global development dataset and space dogs dataset.
This project focuses on applying Deep Reinforcement Learning (RL) algorithms to train agents in two distinct environments: Cartpole and Humanoid. The Cartpole agent learns to balance itself, while the Humanoid agent learns to walk.
Investigation of existing reward functions in the context of question generation with large language models through Reinfocement Learning
A codebase for a full stack react app
This project will help us classify images into a number of classes (clothes) using a CNN model. Additionally, adversarial attacks has been implemented on the model and the effected accuracy has been calculated. This demonstrates how minute changes in the images of training data can significantly impact the model's accuracy
Collection of notebooks on various topics of machine learning pertaining learning basics of machine learning and natural language processing and projects for the application of their concepts.
This project explores movie recommendation systems and evaluation on a dataset of 100 million Netflix customer ratings. It covers recommendation methods including context and content based ones and provides a user-friendly web interface to provide user to enter movie/user id to get recommendations of similar movies.
The Phone Directory project showcases my proficiency in C++ programming, highlighting skills in memory management through pointers, the application of Object-Oriented Programming (OOP) concepts, and the utilization of complex data structures, including trie, hashmaps, and sets.
A feature rich app using React, Redux, and Socket.io for seamless real-time messages with private and group chats.
Tracking emails to know that when the recipient open them
A Demo project made in ReactJS to showcase Optimization done using Server Side Rendering (SSR) via incorporating Next.js module and implementing server code.
A mobile app for the visually impaired includes a Braille keyboard for typing using gestures and buttons, with text-to-speech and haptic feedback. Object Detection module using a camera for real-time awareness through audio feedback. The app uses MobileNetwork with SSD model to classify objects in images
Twitter clone made in Android Studio using Twitter's API in Java. Created a user-friendly interface with features like login, home feed, trending, follow, image uploads to tweets, and advanced options such as long-click actions and dark mode. The project demonstrates proficiency in mobile app development and object-oriented programming.
The primary objective of this project is to investigate the Universal Adversarial Perturbations (UAPs) of image classification models. The project aims to deepen our understanding of adversarial vulnerabilities and proposes measures to fortify the security and fairness aspects of machine learning systems.