Projects
These are some projects that I have worked on, both personal and teams. Find more: GitHub.
Developed a full-stack text summarization system with FastAPI backend for authentication and transformer-based summarization. Implemented user-friendly Streamlit frontend for text input and summary visualization, with comprehensive API documentation.
Developed an end-to-end ML pipeline for sentiment analysis using BERT, orchestrated with Apache Airflow for automated workflow management. Implemented MLflow for experiment tracking and DVC for data versioning, with FastAPI for model deployment.
Developed a Question-Answering (QA) system capable of processing user queries and returning relevant passages from a given corpus as answers. Implemented Generative AI (GPT 2) model to provide direct, concise answers based on retrieved passages, and utilized Streamlit to create a UI to interact with the ML system.
Transformer-VAE combines the power of transformer architecture with physics-based radiative transfer modeling for accurate biophysical parameter retrieval from Sentinel-2 satellite imagery. The model leverages self-attention mechanisms to effectively capture complex relationships in spectral data while maintaining physical consistency through the PROSAIL model to significantly improve feature extraction and parameter estimation.