Data and ML engineer building credit-risk, payments, and analytics projects for fintech.
- Credit-risk scoring and feature-engineering pipelines
- Payments and transaction analytics (batch and streaming)
- End-to-end ML apps: data, model, dashboard, Docker, CI
| Project | What it does | Key tech |
|---|---|---|
| credit-risk-scorecard-engine | WoE/IV binning and a PDO logistic scorecard vs XGBoost on the German Credit dataset, with PSI monitoring and SHAP/LIME | Python, optbinning, XGBoost, SQL, Docker |
| clickhouse-payments-analytics | Live payments-analytics stack: a producer streams events into ClickHouse, materialized-view rollups, a SQL anomaly view, three Superset dashboards | ClickHouse, Superset, Python, Docker |
| payment-retry-ab-test | A/B test analysis of a payment retry strategy: two-proportion z-test, confidence intervals, guardrails, segmentation | Python, scipy, statsmodels, Streamlit |
| ai-ops-workflow-automation-platform | FastAPI service that routes and escalates tickets through workflows and a LangGraph agent with approval gates and RBAC | FastAPI, PostgreSQL, pgvector, Docker |
| pyspark-aml-transaction-analysis | PySpark AML pipeline: window-function features, five typology rules, weighted risk scoring, XGBoost with SHAP and MLflow | PySpark, XGBoost, SHAP, MLflow, Streamlit |
