Enterprise Data Strategy & Decision Systems | Consumer Credit & Risk Analytics | Data Governance & Model Validation

No Cold Handoffs: Logic, controls, validation, evidence, and interpretation travel together.

Senior analytics and applied data science leader with 16+ years of experience across Wells Fargo, the Office of the Comptroller of the Currency, and the U.S. Census Bureau.

I build enterprise data strategies and governed decision systems that connect:

• business and regulatory requirements
• source-data lineage and fitness
• analytical methodology
• decision and treatment logic
• validation and effective challenge
• implementation controls
• monitoring and exception evidence
• executive decision support
• controlled operational follow-through

At Wells Fargo, I was promoted from Vice President to Executive Director while serving as principal analytical architect for Consumer Auto programs exceeding $1B in exposure and customer impact. My experience includes CPI analytical-system ownership from 2018–2025, leadership across 15+ additional remediations, enterprise credit-furnishing governance, and SME or peer-review support for 10+ additional credit-reporting matters.

At the OCC, I supported 16 economists and examination teams with longitudinal consumer-credit data, survival and logistic modeling, CECL and life-of-loan PD analysis, credit benchmarking, alternative-data engineering, and published research support.

My public portfolio demonstrates how those professional disciplines translate into proactive credit strategy, enterprise data governance, model and decision validation, reconciliation, remediation testing, and executive-ready analytical evidence.

LinkedIn Profile · Repository Map · Professional Foundation · Technical Toolkit

I integrate fragmented data, independently managed processes, and cross-functional requirements into governed analytical environments that support consistent execution and reliable executive decisions.

I design transparent rule, strategy, and treatment frameworks with configurable parameters, account-level outcomes, reason codes, monitoring, and archived evidence.

I apply independent recalculation, challenger and benchmark analysis, source-to-target reconciliation, sensitivity testing, UAT, exception management, and effective challenge.

I translate complex analytical and operational evidence into concise narratives covering the issue, facts, risks, dependencies, options, customer and control implications, recommendations, and required decisions.

I connect requirements, methodology, population logic, code, validation evidence, execution artifacts, decision records, controls, and follow-through across independently accountable teams.

This GitHub portfolio is a public proof layer for enterprise data strategy, governed analytics, and decision-system architecture.

The repositories are not isolated notebooks or dashboard demonstrations. Each independently developed system pairs code with some combination of:

• business and analytical requirements
• controlled parameters and strategy configurations
• synthetic, anonymized, or demonstration data
• QA and validation evidence
• reconciliation and exception outputs
• UAT scripts and acceptance checks
• reason-code and decision-path evidence
• run registries and archive tables
• executive summaries and presentations
• technical runbooks and operating guidance

The repository organization reflects the professional documentation and version-control discipline I used at Wells Fargo.

Business Context and Decision Requirements
→ Data Fitness, Lineage, and Reconciliation
→ Model, Policy, or Strategy Development
→ Scenario, Sensitivity, and Challenger Testing
→ Decision or Treatment Execution
→ Financial and Customer-Impact Quantification
→ Validation, Monitoring, and Exception Management
→ Executive Narrative and Governed Follow-Through

The recurring architecture is:

Governed Inputs
→ Transparent Logic
→ Configurable Parameters
→ Account-Level Outcomes
→ Reason Codes and Exceptions
→ Validation Evidence
→ Executive Interpretation
→ Reproducible Handoff

Enterprise Credit Decisioning Strategy Simulator — PostgreSQL

Governed simulator for generated application populations, estimated PD, LGD, Expected Loss, affordability, exposure, counteroffer, review, approval, and decline logic.

Demonstrates:

• deterministic application and risk generation
• configurable product and credit-policy strategies
• baseline and challenger comparison
• matched-population analysis
• affordability and exposure controls
• counteroffer feasibility and governance
• reason-code and decision-path traceability
• strategy-frontier analysis
• archive-backed evidence
• post-campaign acceptance testing

Forensic Data Integrity Gatekeeper — Python

Pre-model diagnostic engine that evaluates raw dataset health before analytical or decision-system use.

Detects and summarizes:

• hidden and disguised nulls
• dominant-value concentration
• zero-variance fields
• format and type defects
• inconsistent categorical values
• anomalous distributions
• field-level and dataset-level severity
• remediation and review priorities

Produces audit ledgers, executive scorecards, and validation-ready handoff evidence.

Enterprise Reconciliation Reporting — SAS

Metadata-driven A/B validation framework for large-scale dataset comparison.

Supports:

• schema-drift diagnostics
• key-only and value-level differences
• absolute and relative numeric tolerances
• date, time, and datetime tolerances
• character normalization
• duplicate-key controls
• severity-ranked exceptions
• run metadata and UAT assertions
• controlled evidence exports

Designed to reduce false positives while preserving reproducible and reviewable validation evidence.

Metro 2 Remediation Sandbox — PostgreSQL

Longitudinal credit-reporting environment for controlled remediation design and testing without PII.

Includes:

• five-million-account baseline portfolio
• month-level performance truth tables
• financial and credit-impact windows
• cure and manual-review logic
• treatment assignment
• delta-based remediated fields
• before-and-after validation
• customer-impact and population evidence
• historical outcome and reason-code traceability

Demonstrates governed remediation testing, temporal state management, and explainable credit-reporting outcomes.

Insurance Coverage Reconciliation — SAS

Metadata-driven interval-reconciliation engine for comparing proof-of-coverage evidence with lender-placed CPI policy windows.

Demonstrates:

• overlapping-interval consolidation
• date and coverage validation
• configurable lapse thresholds
• policy-adjustment factors
• parameter governance
• transparent customer-impact logic
• account-level audit evidence
• repeatable exception review

Treasury-Indexed Remediation Liability Engine — SAS

Memory-efficient financial recalculation engine using one-year Constant Maturity Treasury rates and custom SAS functions.

Technical features include:

• PROC FCMP
• in-memory two-dimensional arrays
• chunked account processing
• annual compounding
• final remainder-interest calculations
• parameterized Treasury-rate application
• account-level recalculation evidence
• executive liability summaries

Demonstrates scalable financial-redress calculation while controlling processing and memory demands.

Survival Retention and Scenario Engine — Python

End-to-end time-to-event modeling and scenario-analysis framework.

Combines:

• K-Means personas
• Kaplan-Meier curves
• Cox proportional hazards
• customer and segment risk analysis
• high-risk scenario simulation
• time-to-event runway
• automated PPTX and PDF reporting
• technical and executive model evidence

Demonstrates survival modeling, scenario-based risk reduction, segmentation, and interpretable analytical reporting.

View the repository

A two-module PostgreSQL decisioning environment demonstrating how generated application data, policy controls, strategy testing, account-level outcomes, validation evidence, and executive interpretation can operate together without exposing PII or proprietary credit policy.

Generates deterministic application populations with:

• product type
• score band
• borrower profile
• requested exposure
• APR proxy
• monthly payment
• payment-to-income ratio
• estimated-PD proxy
• LGD
• Expected Loss

Campaign scope:

• 19 governed scenarios
• 50,000 applications per scenario
• 950,000 archived scenario rows
• product-by-score risk surfaces
• configurable scenario levers
• Expected-Loss impact decomposition

Consumes Module 1 outputs and applies configurable policy and treatment controls to simulate:

• approvals
• counteroffers
• manual reviews
• declines
• approved exposure
• approved Expected Loss
• affordability and exposure constraints
• selective and aggressive counteroffer paths
• baseline and challenger comparisons
• strategy-frontier trade-offs

Campaign scope:

• 39 governed strategy runs
• 50,000 applicants per run
• 1.95 million archived decisions
• 20 matched comparison groups
• seven scenario families
• baseline-relative challenger analysis
• reason-code and path evidence

The repository includes:

• business and analytical requirements
• relational data model
• configurable parameter tables
• strategy definitions
• reason-code logic
• validation and QA scripts
• matched comparison outputs
• run registries
• archive persistence
• executive summaries
• technical runbooks

The simulator demonstrates:

• governed credit-strategy architecture
• pre-production decision testing
• configurable policy and product controls
• account-level reason codes
• matched-population comparison
• Expected Loss and affordability trade-offs
• archive-backed evidence
• challenger-strategy assessment
• counteroffer governance
• executive decision support

It is not a dashboard-only demonstration. It is a complete decision-system environment with relational implementation, controlled parameters, generated samples, validation evidence, archived campaign results, and executive reporting.

The portfolio architecture reflects disciplines developed across three highly regulated public and financial-sector environments.

Wells Fargo | 03/2018–04/2026

• Executive Director – Senior Lead Analytics Consultant | 06/2022–04/2026
• Vice President – Lead Analytics Consultant | 03/2018–06/2022

Promoted while serving as principal analytical architect for Consumer Auto programs exceeding $1B in exposure and customer impact.

• Authored, reviewed, edited, and presented Decision Forum materials covering the issue, supporting facts, risks, dependencies, available options, customer and control implications, recommendations, and required decisions.
• Coordinated input across leadership, Legal, Compliance, Risk, Technology, Credit Bureau Management, Execution, Independent Testing & Validation, Audit, and regulatory stakeholders.
• Ensured approved decisions and changes were formally documented before control review and execution.
• From summer 2024 onward, served as the standing Consumer Auto Finance representative in daily IT&V escalation meetings, coordinating evidence, ownership, resolution, and follow-through.

From 2018 through 2025, designed, operated, and governed critical CPI analytical and execution processes, including:

• proof-of-insurance intake and quality assurance
• Direct and Indirect Auto recalculation end-user computing tools
• CPI coverage adjustments
• customer financial redress
• time-value-of-money loss-of-use relief
• prior-payment netting
• account-level decision-path flags
• manual-review integration
• credit-reporting redress
• recurring 1099-INT and 1099-MISC tax reporting
• weekly production and executive reporting
• complaints, Audit, control, and regulatory support
• annual EUCT documentation, validation, access, and approval requirements

Independent Audit found 100% alignment between retained proof-of-insurance evidence and the corresponding third-party vendor data reviewed.

Led CPI credit-reporting redress for approximately 850,000 accounts through approximately 200 external mass-maintenance submissions, with corresponding internal system-of-record updates and account/borrower-level before-and-after lineage across key Metro 2 fields.

• Created a governed archive of 500+ multi-line-of-business mass-maintenance submissions representing more than 100 million account-month furnishing actions.
• Built a SAS-to-Teradata pipeline to map disparate schemas, standardize and validate line-of-business data, and publish governed datasets.
• Established account-month SQL controls aligning remediation and furnishing history to prevent favorable-state overwrites.
• Led UAT as a distinct validation workstream covering data mappings, decision rules, outputs, and exception handling.
• Separately evaluated 20M+ historical Consumer Auto Finance accounts through the control framework. Most required no update; the evaluation isolated the smaller subset requiring targeted redress.
• Successfully defended a unified correction that accelerated customer relief while preserving accountability reporting identifying whether remediation or furnishing activity caused each overwrite.
• Designed account- and impact-month-specific re-furnishing logic that preserved favorable customer reporting and saved approximately $60,000 per submission.

Led analytical workstreams across 15+ Consumer Auto remediations, including:

• Tier 1 SCRA proactive application-of-benefits matters
• COVID-deferral remediations
• disaster relief
• a major data-center outage
• inaccurate payoff-date reporting
• multiple credit-furnishing matters

Owned the analytical components of the Remediation Target Operating Model:

• Data Analytical Approach
• Population Identification Workbook
• SAS/SQL code
• population outputs
• run procedures
• validation evidence
• Population Identification Review responses
• Decision Forum analysis
• specialized data and review dependencies

For SCRA work, reconstructed Direct Auto account lineage across multiple systems of record and account-recycling processes. Where evidence proved that apparent conflicts resulted from system migrations rather than actual account activity, resolved the discrepancy analytically. Where evidence remained insufficient, developed and escalated risk-mitigation options aligned with SCRA requirements.

Beyond the 15+ remediations led, served as an enterprise furnishing SME, peer reviewer, and mentor on 10+ additional credit-reporting remediations.

Built or supported reusable Consumer Auto processes for:

• governed Common Code
• execution Standard Data Format transformation
• Teradata-based Universal Data Layer archival delivery
• tax-reporting standardization
• business requirements and Job Aids
• operating and run procedures
• department-wide training
• knowledge continuity after staff turnover
• aged IT&V review resolution

Research Analyst, Credit Risk Analysis Division | 03/2014–03/2018

Supported 16 economists and examination teams with model-ready data, analytical methods, validation logic, benchmarks, and decision evidence.

Selected experience includes:

• Developed and tested Cox proportional hazards and discrete-time survival/logistic frameworks for delinquency, default, and prepayment.
• Constructed and validated longitudinal bureau panels spanning 875,516 loans and 2,363,261 loan-year observations.
• Supported borrower, loan, macroeconomic, term-age, prime/subprime, lifetime-risk, and pricing analysis.
• Independently recalculated outputs, tested sensitivities, reconciled source data, and translated results into Tableau views.
• Led review of third-party PD vendor contracts and evaluated data fitness for CECL and life-of-loan PD analysis.
• Built syndicated-credit entity-resolution and fuzzy-matching logic for cross-bank internal risk-rating comparisons.
• Developed obligor-level sequencing logic for first, second, and subsequent delinquency/default events across products.
• Built ArcGIS branch-distance measures for loss-mitigation analysis.
• Developed a Python market-data collector polling stock quotes at one-second intervals for review against macroeconomic events.
• Built a 10,000+ document conversion and SAS ingestion pipeline for fair-housing analysis.
• Provided technical direction and mentoring to junior analysts and interns.

The authors of both the peer-reviewed and OCC long-term auto-loan studies publicly acknowledged me for “excellent research support.”

Selected research supported:

• Risks of Long-Term Auto Loans
• A Puzzle in the Relation Between Risk and Pricing of Long-Term Auto Loans

Survey Statistician | 07/2009–03/2014

Modernized national survey systems supporting the Annual Capital Expenditures Survey and Annual Retail Trade Survey.

Selected experience includes:

• Architected a SAS-orchestrated pipeline refreshing Access tables and JavaScript/HTML dashboards.
• Led a six-analyst nightly production operation and served as escalation point for processing failures, warnings, and unusual data conditions.
• Designed analyst views prioritizing major year-over-year discrepancies and high-impact nonrespondents.
• Automated executive-ready industry narratives, completion summaries, and management reporting.
• Built SAS-to-Excel/VBA validation workbooks for analyst-controlled, on-demand review.
• Developed editing, imputation, response-tracking, outlier, disclosure-avoidance, benchmarking, and documentation controls.
• Maintained methodology, runbooks, troubleshooting guidance, and handoff documentation.
• Translated validated results into management summaries and annual published economic narratives.

SAS Enterprise Guide · SAS Macro Programming · PROC SQL · PROC FCMP · SQL · Teradata · Python · Git/GitHub · Tableau · ArcGIS · Excel/VBA · Microsoft Access · JavaScript/HTML · PuTTY Batch Processing · SharePoint · Jira · PowerPoint

PostgreSQL · pandas · NumPy · scikit-learn · lifelines · matplotlib · openpyxl · python-pptx · ReportLab · Relational Data Modeling · Archive and Evidence Tables

Cox Proportional Hazards · Kaplan-Meier · Discrete-Time Survival and Hazard Models · Binary and Multinomial Logistic Regression · Ordinary Least Squares · K-Means · Probability of Default · Loss Given Default · Expected Loss · CECL · Life-of-Loan PD · Scenario and Sensitivity Analysis · Policy-Rule Simulation · Champion/Challenger Comparison · Strategy-Frontier Analysis

Independent Recalculation · Benchmark and Challenger Analysis · Matched-Population Comparison · Source-to-Target Reconciliation · Data Quality and Lineage Review · UAT and Acceptance Testing · Reasonableness and Sensitivity Testing · Monitoring and Exception Management · Documentation Standards · Effective Challenge · Executive Decision Evidence

Data Standardization · Cross-LOB Integration · Requirements Traceability · Data Analytical Approaches · Population Identification Workbooks · Run Procedures · Governance Checkpoints · Decision Forums · Executive Narratives · Escalation Management · Control Reviews · Reusable Playbooks · Training and Knowledge Transfer

Consumer Credit Risk · Auto Finance · Credit-Bureau and Metro 2 Furnishing · FCRA · CECL · PD/LGD/Expected Loss · CPI · SCRA · Loss Mitigation · Syndicated Credit · Federal Bank Supervision · Remediation and Customer-Impact Analytics

• Bachelor of Arts in Economics, Minor in Mathematics — Virginia Tech, 2005–2009
• SAS Base Programming Certification — 2009
• SAS Advanced Programming Professional Certification — 2009
• Lean Six Sigma Yellow Belt — Office of the Comptroller of the Currency, 2016
• Associate Citation in Project Management — George Washington University, 2013

A system is not complete when the code runs or the output is produced.

A system is complete when the relevant stakeholders can understand:

• what question was being answered
• which data were used
• how the logic operated
• which assumptions and parameters mattered
• how the output was tested
• which controls were applied
• what exceptions remain
• what decision is required
• how the result should be implemented
• how ownership and follow-through will be maintained

That philosophy appears throughout the portfolio:

Data integrity before modeling
Lineage before inference
Validation before reliance
Strategy testing before implementation
Reason codes before unexplained outcomes
Reconciliation before closure
Executive interpretation before action
Documentation before handoff

The objective is not complexity for its own sake. It is analytical work that is transparent, reproducible, governable, and useful.

I also use generative AI as a controlled accelerator for synthesis, narrative development, presentation refinement, and risk-control ideation—while retaining source verification, human judgment, and accountability.

All public repositories use synthetic, anonymized, or demonstration data.

These projects do not expose:

• customer or personally identifiable information
• employer-owned code
• production credit policy
• proprietary remediation rules
• confidential model inputs
• regulated operational pipelines
• internal systems or restricted documentation

The repositories are designed to demonstrate transferable methodology, enterprise architecture, validation discipline, governance, documentation quality, and executive communication in a public setting.

• LinkedIn
• GitHub

Analytics • Decision Systems • Consumer Credit

Logic, controls, validation, evidence, and interpretation travel together.