portfolio
My Projects

PersonalGPT

PersonalGPT is your very own chatbot that answers the questions based on your provided documents and what's more: Ask questions to your documents without an internet connection, using the power of LLMs. 100% private, no data leaves your execution environment at any point. You can ingest documents and ask questions without an internet connection!

Real-Time Stock Market Data Processing Pipeline

Spearheaded the development of an advanced real-time stock market data pipeline. Extracted live data from Finnhub.io and processed over 350,000 data points per hour, reducing Finnhub.io data ingestion latency by 15%. Created a predictive Geometric Brownian Motion stock price model that provided actionable insights into price trends.

Palm&Peaks-Travel Planner

Spearheaded the development of Palm&Peaks, a travel booking application that integrates flight, tour, and hotel bookings into a single platform. Developed a dynamic pricing engine that processes costs and taxes in user-preferred currencies, enhancing user experience and engagement. Successfully implemented data scraping from Yatra.com and Kayak, handling over 500,000 data points daily, streamlining the booking process, and improving data accuracy. Integrated Stripe for payment processing.

A Study on Image Caption Generators

Engineered multiple image caption generators using pre-trained CNNs (VGG16, ResNet50) and RNNs (LSTM, GRU). Evaluated and compared model combinations with BLEU scores on Flickr8k, and MSCOCO to identify the most accurate, meaningful captions.

Text-Summarizer

Designed and built an end-to-end text summarization system using Pegasus model. Implemented a robust data pipeline for efficient data ingestion, evaluation, and transformation. Trained the model and evaluated it to ensure high-quality summaries. Developed a FastAPI web application for easy user interaction. Containerized the application with Docker for seamless deployment.

Drowsiness Detection System

Developed a drowsiness detection system using OpenCV and dlib libraries. Engineered a real-time video processing pipeline that detects facial landmarks and calculates the Eye Aspect Ratio (EAR) to determine drowsiness. Integrated a sound alarm to alert the driver when drowsiness is detected, ensuring road safety.

Do you have a project ?

Full Stack Data ScientistHire ME!