Projects
Code and Pdf's can be found in the GitHub repositories for each project.

Improvising PointNet Architecture for 3D Object Classification

[Jan ’25 - April ’25]

(Prof. Shanmuganathan Raman, IIT Gandhinagar) | Project Link

  • Enhanced PointNet by integrating PMFI (PointMixup-based Multi-scale Feature Interaction) modules and k-NN for improved local-global feature learning.
  • Evaluated advanced variants (xLSTM, KAN, contrastive learning), achieving 90.56% accuracy, surpassing baseline PointNet.
  • Tech Stack: Python, PyTorch, NumPy

Library Management System

[June ’24 - July ’24]

Project Link: Library Management System

  • Designed and implemented a Library Management System in C++ to streamline book management and student interaction.
  • Added features for student registration, book borrowing, and secure login, applying OOP principles.
  • Integrated SQLite database for persistent and efficient storage of student and book records.
  • Tech Stack: C++, Object-Oriented Programming, SQLite

Issue Tracker Portal

[May'24]

(Under Metis Dev Club, IIT Gandhinagar) | project link

  • Developed an Issue Tracker enabling students to get their issues resolved.
  • Tech Stack:HTML, CSS, JavaScript, Bootstrap, Node.Js , MongoDB.
  • Login authentication using password hashing.
  • Adminportal with realtime query resolution.
  • Email conformation to users upon query resolution.

Portfolio Website

[Feb'24]

(Advisor - Created by myself from scratch) | project link

  • Created a website to showcase my projects, resume and other important information.
  • Learned to create a responsive website for desktop as well as for mobile phone.
  • Used Html,CSS, JavaScript to create the website and deployed over GitHub Pages.

Human Activity Recognizer-Machine Learning Project

[Jan'24 to Feb'24]

(Advisor - Professor Nipun Batra, Professor- IIT Gandhinagar) | project link

  • Developed a machine learning model utilizing algorithms such as Decision Trees and Random Forests.Aimed at recognizing human activities (e.g., Walking, Sitting, Laying) from time-series data.
  • Learned featurization and dimension reduction using Tsfel library. Hypertuned the parameters to get the best bias-variance tradeoff.
  • Deployed the model and validated predictions using real acceleration data obtained from smartphone sensors.

Data Narratives

[Jan'23 to April'23]

(Advisor - Professor Shanmughanathan Raman, IIT Gandhinagar) | project link

  • Analyzed given dataset using machine learning libraries like numpy, pandas, matplotlib etc.
  • Preprocessed the dataset to clean outliers and prepare it for analysis.
  • Utilized processed data to analyze trends and behaviors.
  • Applied various machine learning techniques to derive insights from the data.