Projects

Built a CNN model using U-Net architecture for segmenting satellite images using TensorFlow and Keras. Deployed on Hugging Face for real time prediction.

Developed as part of Llama-Impact Hackathon. Solves the problem of using government websites easier for common people, navigating and providing legal advice.

A forecasting pipeline that retrieves real-time Bitcoin price data via API, preprocesses it, and trains an LSTM model to predict future price trends. Orchestrated with Apache Airflow.

End-to-end ML pipeline for identifying kidney tumors from CT scan images. Includes data ingestion, model training, evaluation, and deployment with MLOps principles.

An application to compare the performance and accuracy of SQL Agent and RAG for tabular data. SQL Agent was suitable for tabular data because of its precise answers.

Built an end-to-end ML pipeline to predict surge pricing for Uber and Lyft rides using historical ride, weather, and event data. Leveraged MLOps principles with MLflow, DVC, and Docker.

Built a CNN for detecting faces using TensorFlow and Keras. Augmentation on images using Albumentations.

Web-based application to streamline and automate airline core operations: flights, passengers, bookings, employee details, routes, fares, and transactions.

An agentic AI career guidance system using LangGraph and multi-agent orchestration to provide personalized career recommendations, skill gap analysis, and learning path generation.

An AI-powered travel planning assistant using multi-agent workflows to generate personalized itineraries, recommend destinations, optimize budgets, and provide real-time travel insights.