Projects

Image Style Transfer
Inspired by the paper “Image Style Transfer Using Convolutional Neural Networks” (Gatys et al.), I implemented neural style transfer from scratch using PyTorch. This included building core components such as content loss, style loss with Gram matrices, preprocessing/postprocessing, and an optimization loop (L-BFGS). Initially attempted with NumPy + ONNX, but later refactored to PyTorch for autograd support. The project demonstrates how artistic styles (like Van Gogh’s Starry Night) can be blended with content images to generate visually compelling results.

House Price Prediction
This project demonstrates the application of machine learning techniques to predict house prices based on various features. By analyzing the dataset, preprocessing the data, and selecting an appropriate model, we were able to achieve a high level of accuracy in predicting house prices. The trained model can be further refined and deployed.

8-bit CPU
This project features the design and implementation of a custom 8-bit CPU. It includes key components such as a custom instruction set, ALU, registers, memory, and a control unit, all working together for basic computation tasks. The CPU is built using Verilog.

Chat Bot
This chatbot project demonstrates the application of AI and ML techniques for natural language processing tasks. By training on a dataset of intents and responses, the chatbot is able to understand user queries and provide appropriate responses, making it a useful tool for various applications, including customer support, and more.

Movie Recommendation System
A movie recommendation system, is an AI/ML-based approach to filtering or predicting the users’ film preferences based on their past choices and behavior. It’s an advanced filtration mechanism that predicts the possible movie choices of the concerned user and their preferences towards a domain-specific item, aka movie.

Contact Sphere
Built a Contact Managing web app which allow user to store and access their family or friends contact info from anywhere. Used JAVA(Spring Boot) as Backend, Thymeleaf for Frontend and Postgres for Database

Tic Tac Toe
Android based tic tac toe game app