Data Scientist · Analytics Engineer · Agentic AI Researcher
I build analytics and data-science systems that turn messy operational, geospatial, financial, and document data into decision-ready workflows. My recent work focuses on agentic AI evaluation, temporal leakage, financial-risk benchmarks, and practical BI/data products.
Repository material supporting the Azure AI Document Intelligence and Power Platform case study, focused on scanned-report extraction, validation, and reporting workflows.
Fraud-detection modeling work around tabular transaction data, feature preparation, and classification-oriented experimentation.
Applied regression project for house-price prediction, useful as a compact example of exploratory analysis, modeling, and error evaluation.
Designed an OCR-driven reporting workflow using Azure AI Document Intelligence, Power BI, Power Apps, and Power Automate to replace manual document processing.
Built a low-cost cloud routing workflow combining ArcGIS, Azure Functions, Blob Storage, Power Automate, and Power BI to automate daily geo-routing operations.
Private multi-agent analytics platform for stock research, deterministic diagnostics, forecasting comparisons, backtesting, and opportunity scoring.
Dissertation research plan connecting temporal leakage, counterfactual replay, herding, liquidity fragility, and systemic risk in AI-mediated financial markets.
Benchmark design work for AI-mediated financial markets, market-quality harm, manipulation-risk indicators, and systemic-risk signals.
Research in progress on leakage-aware counterfactual replay for financial AI agents, decision reversal, and risk-evaluation validity.