Explore a curated selection of my most recent projects that highlight my expertise in artificial intelligence, computer vision, machine learning, and data analytics.

Document QnA App

RAG Pipeline

A memory-infused question-answering system utilizing OpenAI's Large Language Models to intelligently process documents in formats like PDF, DOCX, and TXT. It features an interactive Streamlit interface for user interaction, advanced document handling capabilities, and efficient, context-aware response generation through OpenAI Embeddings. Deployed with Docker and Heroku, it ensures consistent performance and scalability.

Technologies Used:

OpenAI's GPT 3.5 Model, OpenAI Embedding Model, Chroma, Docker, Heroku, Streamlit

Other Projects

Reddit Comments Classifier

Fine-tuned a BERT model to classify Reddit comments into three categories for a medical use-case, and deployed it using Streamlit.

Technologies: Python, PyTorch, Transformer, NLP

Database Modeling in PostgreSQL

Designed and implemented data models in PostgreSQL to simplify complex schemas and improve query efficiency.

Technologies: PostgreSQL, Python

Food Delivery Analytics

Analyzed trends in a food delivery business to optimize data-driven decisions through detailed analytics and visualization.

Technologies: SQL, Microsoft PowerBI

Data Acquisition from an MQTT Broker

Implemented an efficient MQTT broker connection to stream process variables from remote IoT devices on oil and gas vessels.

Technologies: Python, AWS Lambda, JSON, Paho-MQTT

Loan Default Prediction

Developed a cloud-based solution to predict loan defaults, comparing models built with AutoML and custom hyperparameter tuning.

Technologies: Python, Azure AutoML, Azure SQL Database

People Counter App

Developed a smart video IoT solution that counts people in a video frame, enhancing surveillance and monitoring applications.

Technologies: Python, Intel® OpenVINO™ Toolkit, OpenCV, Node.js, npm, FFmpeg, MQTT

Find me on ...