Recent Projects
A sample of recent engagements across model development, CECL, IRR/ALM, model validation, and model risk governance — at banks and credit unions ranging from $20 billion to $75 billion in assets.
CECL & Stress Testing Model Suite — $75B Bank
Built a complete loss forecasting model suite covering every major consumer and wholesale portfolio for a $75 billion bank — designed for both CECL and stress testing — from raw data through production.
- Mapped merger history and built a unified developmental data warehouse with automated quarterly ETL for CECL and stress testing inputs
- Developed, implemented, and tested loss forecasting models for every major portfolio — consumer and wholesale — in a leading CECL platform
- Built an independent parallel process in R to verify all model forecasts — a standard Spero practice on every development engagement
- Delivered a model suite used for both CECL reserve estimation and stress testing scenarios
Loss Forecasting for Run-Off Consumer Portfolios — >$25B Credit Union
Built custom CECL loss forecasting models for a large credit union’s run-off portfolios — solar panel loans, home improvement loans, and refinanced student loans — where standard approaches and sample histories were inadequate.
- Designed bespoke loss forecasting frameworks for three run-off portfolios with short sample histories and no off-the-shelf methodology
- Built custom implementations with overlay adjustment capability to handle data limitations specific to run-off books
- Delivered models maintainable and updatable by client staff
Risk Rating Model Implementation — >$20B Credit Union
Implemented a risk rating model in a leading third-party platform for a credit union exceeding $20 billion in assets, delivering a production-ready system with enhanced portfolio risk assessment capability.
- Implemented risk rating model in a leading third-party platform on behalf of the credit union’s loan portfolio modeling department
- Validated model logic and output against expected behavior across the portfolio
- Documented implementation to support ongoing use and regulatory review
Ongoing Monitoring & Scenario Analysis — $75B Bank
Designed and documented a complete ongoing monitoring framework for a $75 billion bank, including backtesting, a parallel forecasting process, and policies and procedures for regulatory compliance.
- Backtested model results to resolve audit findings and establish a defensible performance record
- Built ongoing monitoring package and parallel statistical forecasting process for legacy models running in a third-party platform
- Wrote ongoing monitoring policies and procedures suitable for examiner review
Model Risk Management Function Review — $35B Credit Union
Assessed and rebuilt a $35 billion credit union’s model risk management function through gap analysis, policy redesign, and direct participation in hiring senior MRM staff.
- Conducted gap analysis and stakeholder interviews to assess MRM function against SR 11-7 expectations
- Recommended structural and procedural changes to policies and governance documentation
- Participated directly in hiring the institution’s new MRM manager — an engagement that went beyond advisory
CECL Platform Migration — >$30B Credit Union
Led complete CECL platform migrations including data, model rebuilds, and documentation for a credit union exceeding $30 billion — from Sageworks to Moody’s ECL and separately to SAS ECL.
- Managed full platform transitions including data mapping, model rebuilds, and parallel validation processes
- Supported migrations to Moody’s ECL and SAS ECL across multiple credit union clients
- Delivered production-ready model suites and documentation at each transition — not just migration support
CECL Regulatory Documentation Package — >$25B Credit Union
Rebuilt the complete regulatory documentation suite for a credit union exceeding $25 billion following an internal decision to transition CECL platforms — policies, process narratives, methodology summaries, and production control workbooks.
- Updated CECL policy, ACL process narrative, and methodology summary to reflect new model classifications and data sources across all retail and mortgage portfolios
- Redesigned production quality control workbook to reduce manual oversight while keeping it maintainable by non-technical staff in Excel
- Standardized end-to-end process flow diagrams for all active models, each with dedicated control and checkpoint pages
Data Infrastructure Migration & Process Automation — $35B Credit Union
Migrated a $35 billion credit union’s CECL analytical infrastructure to Snowflake and automated previously manual monthly workflows across consumer portfolios.
- Consolidated auto, credit card, and personal loan portfolio processes into a single unified automated workflow on Snowflake
- Automated monthly recovery tracking in Python, eliminating manual data transfer between model outputs and reporting charts
- Delivered a production-grade pipeline that reduced operational risk and accelerated the monthly CECL cycle
Qualitative Adjustment Framework Redesign — >$25B Credit Union
Identified critical flaws in a credit union’s CECL qualitative adjustment process — including a questionnaire failing to elicit usable LOB responses and unlikely to survive regulatory scrutiny — and rebuilt the framework from the ground up.
- Observed live CECL production runs to surface gaps in data acquisition, quality validation, forecast configuration, and output storage
- Identified and documented missing steps and ambiguous procedures to ensure executability by staff unfamiliar with the process
- Redesigned the LOB qualitative questionnaire to supplement narrative responses with quantitative metrics, transforming a vulnerable process into a defensible, documented framework
Validated QRM Interest Rate Risk Model — >$25B Credit Union
Independent validation of a major QRM model update, confirming the accuracy of NII, EVE, and rate sensitivity estimates across retail portfolios and deposits.
- Validated changes in NII, EVE, and associated sensitivity analysis for retail portfolios and deposits in a QRM/IRRBB framework
- Independently confirmed rate sensitivity calculations across assets and liabilities following a significant model update
- Delivered findings and documentation suitable for regulatory review