Reporting brought in-house at Morgan & Morgan
Data products / analytics / applied AI
I build data products that turn raw systems into clear decisions.
I design analytics, pipelines, models, and AI products that hold up beyond the demo. The goal is simple: make complex data useful to the people doing the work.
Previously three to five days
Governed compliance and compensation pipelines
Hybrid PTO approach adopted by leadership
Selected live work
Built, tested, and available to explore.
A cross-section of deployed work across data engineering, analysis, data science, and AI engineering.
Browse all 12 live projectsFinance & risk
Payments Fraud Risk Data Platform
A governed fraud-risk validation register that publishes all 1,852,394 allowlisted simulated events for bounded analytical queries while keeping identity-like fields, model scores, and payment actions out of the public system.FocusData Engineering
ResultRetain the ordinary logistic baseline for the measured 1% review queue. It reaches 0.160 PR-AUC and 51.3% recall while the class-weighted challenger performs worse on ranking, recall, and probability error.
PythonPostgreSQLscikit-learnSupabaseFastAPINext.js
Healthcare / Operations
GP Access Planner
A public-data planning product that forecasts recorded general-practice appointments across England while keeping observed access signals and hypothetical capacity explicitly separate.FocusSenior Healthcare Analyst
ResultDelivered a source-traceable 7, 14, and 28-day planning surface for 104 sub-ICBs, backed by 32.9 million validated source rows, rolling-origin evaluation, immutable releases, and a live edge API.
PythonPostgreSQLdbtscikit-learnLightGBMCatBoost
FocusAI engineer / Data engineer
ResultConnects documents, people, events, and citations in one inspectable investigation workflow.
PythonFlaskLangChainsentence-transformersONNX RuntimePostgreSQL + pgvector
Finance & risk
Automobile-Loan First-EMI Default Strategy Portfolio
A retrospective credit-policy platform over 233,154 Indian vehicle loans: portfolio and vintage reporting, score-decile and segment risk analytics, a policy workbench with editable economics, a per-loan inspector, and post-deployment monitoring.FocusData Analyst
ResultA credit-policy analyst can test where to draw a first-EMI risk line, see the confidence interval around the answer and who it declines, and take a recommendation or a refusal to governance. On the published assumptions the honest output is a refusal: no evaluated band clears zero.
PythonJavaScriptscikit-learnSQLiteFastAPIReact
How I work
Build the path from question to action.
I define the decision, validate the data, compare a useful baseline, and ship the result in a form someone can actually operate.
Name the decision
Who acts, what changes, and what a wrong answer costs.
Earn the dataset
Reconcile definitions, permissions, leakage, and failure paths before tuning anything.
Set the stopping rule
Choose baselines, capacity, uncertainty, and refusal gates before reading the result.
Ship for review
Put the recommendation, explanation, and override in the same operating workflow.
Preferred delivery stack
From first brief to monitored decision system.
This is the complete toolkit I would reach for to shape, build, test, deploy, observe, and report on a data product. It is a preferred operating stack—not a usage counter or a dump of project dependencies.
Product direction
Briefs, user flows, architecture decisions, and an owned backlog before implementation starts.
Application layer
Accessible interfaces, typed contracts, production APIs, and clear client–server boundaries.
Data foundation
Transactional storage, analytical models, orchestration, event streams, and low-latency access.
Analysis and modeling
Reproducible exploration, feature engineering, model training, evaluation, and explainable outputs.
Applied AI
Grounded generation, retrieval, graph context, model access, and portable inference paths.
Quality and security
Unit, integration, browser, lint, dependency, and container checks before anything is released.
Cloud and infrastructure
Portable services, declarative infrastructure, managed data, edge delivery, and environment parity.
Release and operations
Version control, automated delivery, production hosting, traces, errors, and operating feedback.
BI and decision support
Governed metrics, semantic reporting, executive dashboards, ad-hoc analysis, and visual diagnosis.
Career highlights
A short record of outcomes, not a public resume.
As a Data Analyst at Morgan & Morgan, P.A., I build shared data systems, reporting products, and decision models. The profile keeps the public story focused on a few measurable changes.
View achievement snapshotsStart a conversation