Projects

Text Generator based on next character prediction (Machine Learning Project)

[Mar '24]

(Advisor - Professor Nipun Batra, Professor- IIT Gandhinagar) | Project Link | Application link

• This project involves generating text by predicting the next character based on the last 'k' characters. The model used is a neural network with an embedding layer followed by 2 hidden layers. The model is able to capture the semantics of English language and generate meaningful words and phrases.
• Observed the difference in results based on the choice of embedding and architecture of the neural network. Performed hyperparameter tuning to get the best results. Learned to deploy an application on Streamlit.

Facebook graph clustering based on density (Data Science Project)

[Jan '24]

(Advisor - Professor Anirban Dasgupta, Professor- IIT Gandhinagar) | Project Link

• Implemented a greedy algorithm to extract the densest subgraph from Facebook's graph (with pages being nodes and mutual links being edges). Peeled the 5 densest subgraphs (and hence, detected 5 closely knit communities in the graph) and analyzed their entropy.
• Learned to employ data clustering techniques on large real-world datasets, involving logic and algorithmic optimization in terms of both space and time complexity. Developed using Python and its relevant libraries like NumPy and Pandas.

Human Activity Recognizer (using Machine Learning)

[Jan '24]

(Advisor - Professor Nipun Batra, Professor- IIT Gandhinagar) | Project Link

• Made a machine learning model that uses only decision trees to recognize 6 different human activities by utilizing time series data of acceleration involved.
• Learned to handle time series data by employing featurization and dimension reduction. Performed hyperparameter tuning to select the best model based on the bias-variance tradeoff. Deployed the model and tested the predictions on real acceleration data collected from a smartphone.

Spacecraft Charge Distribution Modelling

[Sept '23 - Nov '23]

(Advisor - Professor Soumyabrata Chakrabarty, Scientist- ISRO SAC, Professor- IIT Gandhinagar) | Project Link

• Conceptualized and executed a project focused on mathematically modeling the charge distribution on the surface of a spacecraft in outer space orbits, with specific emphasis on mitigating risks associated with uneven charge distribution.
• Developed and implemented a methodology utilizing the method of moments to derive equations for charge density on the spacecraft's surface. Employed numerical methods such as Gaussian elimination, Gauss-Jordan algorithm, and TDMA to solve the system of equations, providing a detailed and accurate charge distribution profile. Programming of all mathematical equations and visualizations are done using Python and it's libraries.

Child Safety Monitoring Application, IoT Powered bicycle safety system

[Sept '23 - Nov '23]

(Advisor - Professor Nithin V. George, IIT Gandhinagar) | Project Link

• Developed a comprehensive Android/iOS mobile application, transforming the phone into a smart IoT embedded system for monitoring child safety during cycling.
• Implemented advanced features, including real-time GPS, accelerometer, and gyroscope data monitoring, alongside boundary detection, fall detection, and over-speed alerts.
• Integrated a custom-built mini Google Maps for the IITGN campus, allowing seamless real-time location sharing and communication between the parent and child devices. Utilized MATLAB for backend processing, Simulink for UI design, and TCP/IP protocols for efficient real-time data transfer and voice communication.

Intelligent Game Engine (DSA Project)

[Sept '23 - Nov '23]

(Advisor - Professor Balagopal Komarath, IIT Gandhinagar) | Project Link

• Developed a repository featuring intelligent game implementations, where the computer employs optimal strategies to make informed moves, showcasing proficiency in algorithmic decision-making, especially graphs and path-finding algorithms.
• Engineered a cubesolver capable of generating the most efficient solutions for any 2x2x2 Rubik's Cube, highlighting problem-solving skills and a deep understanding of algorithmic optimization. Additionally, designed a Sudoku solver demonstrating expertise in logical puzzle-solving algorithms.

IITGN Technical Summit's Official Website

[Sept '23 - Nov '23]

(Technical Team, Amalthea, IITGN Technical Summit) | Project Link

• Played a pivotal role as a key contributor in the development team for the official website of Amalthea '23, the annual technical summit at IIT Gandhinagar.
• Made significant contributions to the website's design, functionality, and overall development, ensuring a polished and user-friendly online platform for the summit. Worked collaboratively to implement innovative features and enhance the digital experience for participants and attendees.

Student Resume Generator Website

[Dec '23 - Ongoing]

(Technical Team, Career Development Services (CDS), IIT Gandhinagar) | Project Link

• Developing an innovative website dedicated to the automatic generation of student resumes, adhering to the specific format mandated by the Placement Cell at IIT Gandhinagar.
• Engineered an intuitive user interface enabling students to input their details seamlessly, with the system autonomously generating professional resumes. Emphasized precision and conformity to Placement Cell standards in both design and coding, streamlining the resume creation process for students.

Machine Learning Blogs Website

[Aug '23 - Ongoing]

(Advisor - Professor Nipun Batra, IIT Gandhinagar) | Project Link

• Crafting technically in-depth blogs to unravel the mathematical foundations driving machine learning algorithms, enhancing comprehension of their operational mechanisms and predictive capabilities.
• Visualizing the mathematics behind machine learning algorithms and explaining their working.

Forcasting Stock Prices using Machine Learning

[Jan '23]

(3rd position in HackRush, IITGN's annual hackathon) | Project Link

• Orchestrated a machine learning project centered on predicting stock prices by harnessing the capabilities of the XGBoost algorithm, a powerful ensemble learning technique.
• Engineered a sophisticated predictive model that delved into historical stock data, leveraging XGBoost's algorithmic prowess to enhance accuracy and responsiveness in forecasting. The project showcased adept algorithmic implementation and optimization strategies to navigate the intricacies of stock market dynamics effectively.

Data Narratives

[Jan '23 - Apr '23]

(Advisor - Professor Shanmughanathan Raman, IIT Gandhinagar) | Project Link

• Analyzed and organized the data of highly complex datasets using Numpy, Pandas, Matplotlib and Python.
• Extracted quality information from the dataset and used it to predict trends by using machine learning techniques.

Text Chat Web Application, Metis IIT Gandhinagar

[May '23]

(Advisor - Metis, Web and App Development Club, IIT Gandhinagar) | Project Link

• Deployed a web application that allows users to communicate in real-time.
• Learned the use of NodeJS and SocketIO and creating a large codebase of organized codes.

MetaIITGN

[Jun '23 - Ongoing]

(Technical Council, IIT Gandhinagar) | Project Link

• One of the founding contributors of this platform. Designed web pages using MKDocs and merged them on github.
• Working on making the platform open to discussions adding additional functionalities.