Vaibhav Srivastava
SDE Intern
August 2024 - Present
Made RESTful APIs
Created interactive UIs
Integrated LLMs
Worked on large scale projects
Full Stack Developer
Nov 2023 - March 2024
Next.js, Flask, SQL, Typescript, Python
Click on Card to Learn more >
Live Link
Link to Code
A real-time chat app with React, Firebase, and WebSockets for multiple users, handling frequent messages. Integrated a CNN-based TensorFlow model for Indian Sign Language translation with high accuracy. Enhanced voice message clarity using Large Language Models, aiding people with speech disabilities.
Click on Card to Learn more >
Live Link
Link to Code
An AI-powered app that enhances in-store shopping by allowing users to chat with the app for real-time product information, ingredient details, and safety advice. Users can compare products by taking photos and receive instant insights to make informed decisions. The app also streamlines checkout by generating bills instantly and supporting multiple payment methods, eliminating the need to wait in line.
Click on Card to Learn more >
Live Link
Link to Code
An ADHD-friendly Book Reader providing chapter-wise summaries for a wide range of books from Project Gutenberg. Enhanced readability using formats like Bionic Reading and created flashcard-style summaries with expandable details to minimize information loss. Included features like an AI-powered debate section, a tool that converts highlights into detailed notes, and a dedicated area for personal annotations.
Click on Card to Learn more >
Live Link
Link to Code
A full-stack website, 'Living Lores,' using Next.js 14 and MongoDB, offering a dedicated platform for aspiring story writers. Seamlessly integrating front-end and back-end technologies, the project provides a dynamic space for creative expression and storytelling.
Click on Card to Learn more >
Live Link
Link to Code
An application built with React that facilitates the design of 3D T-shirt models, offering users the ability to customise colours, upload personalised images, and transform their creative ideas into tangible designs.
Click on Card to Learn more >
Live Link
Link to Code
A sleek SaaS landing page with Next.js and Tailwind CSS, enriched with React Tab for seamless user interaction. Demonstrated expertise in modern front-end development for an intuitive and dynamic web experience
Click on Card to Learn more >
Link to Code
Shakespeare GPT," "Gita GPT," and other specialized AI models, fine-tuned with QLoRA on robust, domain-specific datasets. Utilized low-rank adaptation, mixed-precision training, and gradient checkpointing for performance optimization and reduced computational overhead. "Shakespeare GPT" offers context-aware Shakespearean responses, while "Gita GPT" delivers nuanced insights into the Bhagavad Gita. These models showcase advanced contextual understanding and stylistic precision, available on my Hugging Face repository.
Click on Card to Learn more >
Link to Code
A machine learning model predicting video gaming-induced anxiety using various classifiers: k-NN, SVM, decision tree, random forest, and XGBoost. Showcased versatility in model selection for addressing psychological well-being in the tech domain.
Click on Card to Learn more >
Link to Code
A driver drowsiness detection model utilizing CNN and computer vision. Implemented a robust system to identify drowsiness, providing timely alerts to enhance road safety
Click on Card to Learn more >
Link to Code
A machine learning model predicting video gaming-induced anxiety using various classifiers: k-NN, SVM, decision tree, random forest, and XGBoost. Showcased versatility in model selection for addressing psychological well-being in the tech domain.