about

Data scientist specializing in internal audit, fraud examination, risk management, and financial planning & analysis — backed by twenty years of quantitative work.

Since 2005, across construction, manufacturing, retail, and professional services, I’ve delivered fraud detection clustering models, Python-based purchase-to-payment continuous monitoring, ARIMA and SARIMAX revenue forecasts, Time Series Analysis, causal inference on workforce surveys, enterprise ETL pipelines, and board-level analytics — all before the title existed.

LinkedIn GitHub Contact Me

CIA · CFE · DATA SCIENTIST

The Berkeley MIDS from the School of Information formalized that practice: statistics, ML, data engineering, and causal inference assembled into a coherent technical foundation.

I combine both with a background in FP&A, internal audit, and fraud examination to build ML, GenAI, and statistical inference solutions.

Arabic calligraphy is my parallel practice—a discipline that demands full presence, and offers in return a quiet far from the noise of numbers.

‘Those truly fear Allah, among His Servants, who have knowledge’ — Fatir:28

What I Do

The problems I find most compelling are those where domain fluency dictates the model’s architecture, not just its tuning. In fraud examination, understanding the mechanics of a disbursement scheme determines which features must be engineered. In FP&A, knowing how a budget is structurally assembled determines which residuals actually deserve attention.

My background in audit shapes how I approach model validation: a clean reconciliation still needs an explanation. I won’t deploy until I understand what a model actually learned, not just that it passed its metrics.

Early in my career, I built an end-to-end solution spanning data ingestion, EDA, and descriptive modeling through to board-level reporting — visualizations and dynamic reports translated into language executives could act on. The job title said Audit. The work was data science before it had a name.

I operate most effectively at the intersection of domain expertise and quantitative rigor — where the business problem shapes the equation, and the equation refines the question.