Yohan M Markose

The project integrates data from Free Company Dataset (Snowflake) combining with yfinance to fetch financial side of data for available companies. Along with MCP-integration of google maps to figure out the rich details on the ideal or potential locations for the user. These reports are supported by Q&A agents along with VC investment quarterly reports to address and serve as a platform that will provide users with a base for questing and analyzing his interest in specified business domains, company size parameters, and location preferences to generate tailored recommendations and insights.

The project focuses on designing an AI-powered information retrieval system that automates the extraction, processing, and analysis of financial reports. Using Apache Airflow, the system orchestrates the retrieval of NVIDIA's quarterly financial statements, processes them with advanced text extraction methods, and stores the data efficiently for querying. The project incorporates multiple chunking techniques to optimize data segmentation and embeddings to enhance retrieval performance. A Streamlit-based UI enables users to explore the dataset and upload their own documents for analysis, ensuring an interactive and customizable experience.

Conducted a comprehensive Data Science analysis examining the environmental impact of Electric Vehicle (EV) adoption in China, accounting for both road emissions reduction and battery production emissions. The analysis used Bayesian Statistics and probabilistic modeling to simulate real-world scenarios based on historical data. Key Components: Developed statistical models to analyze historical data and identify critical transition points in emissions patterns Conducted deep-dive analysis of identified turning points to understand key contributing factors Created forecasting models using Bayesian statistics to project future emission trends and simulate long-term environmental impact scenarios.