Decision-ready case studies across analytics, engineering, and machine learning.

43.8% Revenue Leakage Found

InkImpact

I built a decision-ready retention analysis that exposed a 43.8% revenue leak, identified high-LTV acquisition channels, and guided budget reallocation toward stronger ROI.

XGBoost + SHAP model explains which channels actually drive conversions

Marketing-Driven Sales Prediction

I built a decision-ready marketing forecasting workflow that used explainable machine learning to predict conversions and clarify which channels were most likely to drive sales.

7.7M Records Across 49 States

SafeRoute Analytics

I built a decision-ready safety analytics system by engineering 7.7M crash records into risk clusters, severity models, and an interactive advisory tool for logistics and insurance use cases.

NLP + ML system predicting booking completions for British Airways

SkyInsight: British Airways Customer Intelligence

I built a decision-ready customer intelligence workflow for British Airways that combined NLP and predictive modeling to explain sentiment drivers and forecast booking completion.

SQL/Python pipeline cut operational reporting time by 40%

E-Commerce Pipeline & Market Analytics

I built a decision-ready retail analytics pipeline that automated reporting, improved forecasting, and gave a founder-operator cleaner visibility into inventory and demand.

Owned Retrieval Pipeline Architecture

Deep Learning RAG Agent

I built the decision-ready retrieval backbone for a deep learning RAG agent, engineering ingestion, vector search, and orchestration so users could get grounded answers with source context.

Hybrid NLP + Anomaly Detection

Guardian Recruit

I built a decision-ready fraud detection system that combines BERT NLP, anomaly detection, and SHAP explainability to flag suspicious job listings with interpretable risk signals.

3 attack vectors quantified — Backdoor, SYN-Flood & Cryptojacking

Lucid ChargeGuard: EV Cyber-Risk Intelligence

I built a decision-ready cyber-risk analysis for EV infrastructure by using SQL, BigQuery, and Looker Studio to quantify attack exposure and prioritize mitigation.