Projects

SQL Query Generator with Gemini

Developed a Streamlit app using Google's Gemini model to translate natural language questions into SQL queries and retrieve data from an SQLite database. The project also includes a script to set up and populate the SQLite database with sample data.

YouTube Transcript Summarizer

Built a Streamlit app to summarize YouTube video transcripts using Google Gemini, converting detailed transcripts into concise summaries.

Chest Xray Classification

Built a comprehensive pipeline using CNN with PyTorch, Docker for containerization, and BentoML for deployment, with data hosted on Amazon S3.

Gemini Nutrition Analyzer

Built a Streamlit app using Google's Gemini Pro Vision model to analyze food images, provide calorie counts, nutritional details.

Multilanguage Invoice Extractor

Built a Streamlit app using Google's Gemini Pro Vision model to perform advanced multimodal analysis of invoice images, providing detailed insights based on the extracted content.

Smart ATS Evaluator

Created a Streamlit app using Google's Gemini models to analyze resumes, providing match percentages, missing keywords, and professional evaluations based on job descriptions.

Calories Advisor

Built a Streamlit app utilizing Google's Gemini Vision model for multimodal analysis of food images, delivering detailed nutritional profiles and caloric assessments.

Gemini PDF Chatbot

Created a Streamlit app that extracts text from PDFs, indexes it with FAISS, and uses Google’s Gemini model to answer user questions about the PDF content.

Gemini AI Suite

Created a suite of Streamlit apps using Gemini models for Q&A, conversational chat, and image analysis.

Correlating Twitter Sentiment with Stock Prices through NLP

Collected 150,000 tweets on Nestle, Dabur, and Britannia from 2010 to 2023 to investigate the correlation between sentiment & stock prices.

Text Summarizer

Built and deployed a Pegasus transformer-based text summarizer web app on AWS.

Blog Generation

A blog-generating web app using Large Language Models, customi by topic, word count, and the user's role—researcher, data scientist, or general writer.

Chicken Disease Classification

A deep learning project implemented as a web application and deployed on AWS. It takes an image of a chicken as input and provides the potential disease diagnosis as output.