Overview :
A machine learning project focused on classifying earthquake events in California and Los Angeles into multiple event categories using seismic data. The project covers data exploration, feature engineering, model comparison, and performance evaluation.

Overview:
An end-to-end machine learning project for classifying EEG recordings into four clinical categories: healthy, generalized seizure, focal seizure, and seizure event. The project combines exploratory analysis, model optimization, ensemble learning, and REST API deployment.

Overview
A machine learning project focused on predicting whether bank clients are likely to subscribe to a term deposit. The project combines exploratory data analysis, feature engineering, class-imbalance handling, and classification modeling to identify the key factors influencing customer decisions.

Orbie — LangGraph Multi-Agent Assistant 

Overview
An intelligent conversational assistant built with LangGraph that dynamically routes user requests to the appropriate workflow: general conversation, PDF-based question answering, or web search. The application integrates LLM reasoning, RAG, document retrieval, and external search into a unified Chainlit interface.

What it serves
Orbie provides a flexible assistant capable of answering general questions, extracting information from uploaded documents, and retrieving up-to-date information from the web. The project demonstrates agentic workflow orchestration, state management, intent classification, and session-based conversation handling.

Orbie — AI Analytics Agent  

Overview

Orbie AI Analytics Agent is a conversational analytics solution that allows users to explore business data using natural-language questions. It serves as an interactive layer between business users and structured data, enabling them to retrieve performance metrics, analyze trends, compare regions or products, and generate visual insights without writing SQL queries.

What it serves

It serves as an accessible analytics layer for business users who need to explore sales, revenue, profitability, customer, product, and regional performance without writing SQL. The solution helps reduce the time required to retrieve and interpret business information while keeping the analytical workflow structured and reliable.