Established Rhyme’s data platform serving internal BI, external Data Products, and AI/ML solutions• Set the strategic vision, road map, and hiring plan• Democratized data access and enabled self-service analytics• Single source of truth for all analytics and AI/MLDirectly supervised 5 individual contributors (3 full-time, 2 contractors)Cloud cost optimization – reduced costs by $80k/yearOriginated first revenue generating analytics productDeveloped Rhyme’s LiveAuth Score the first reliable metric of Prior Auth efficiency at network scaleCatalyzed a full transactional system redesign to unlock the next gen of analytics capabilitiesMaintained relationships with cloud vendor account teamsFostered culture of data excellence within the organizationGenerative AI research and experimentation• Attended multiple AWS re:Invent workshops• Hands-on experience with AWS Bedrock, Prompt Engineering, and Retrieval Augmented Generation (RAG)

Consultation with senior leadership• Presented road map to transition from bespoke engagements to scalable ML operations• Clarified tangible AI/ML capabilities vs hype and misconceptions• Worked with CEO to define Gradient AI platform product vision• Worked with GM on legal settlement product conceptualizationLed team of 5 to implement an inflation adjustment module. Improved model accuracy by over 50%• Catapulted performance from well below humans to outperforming expertsIdentified misalignment of expectations vs capabilities for underwriting product. Successfully recast product as a pricing assistant toolLed adoption of Amazon Sagemaker for ML OpsDrove business impact and ROI analyses. Revamped and standardized customer-facing exhibitsDeveloped models for underwriting and claims in P&C lines: GL, BOP, Personal & Commercial Auto• Uncovered statistical issue negatively impacting all claims models. Solved with thought-leading time-stratified sampling technique. Improved model accuracy by 18%• Developed modeling paradigm to reduce number of model deployments from 6+ to just 1 per engagement• Developed probabilistic confidence interval models, improvement over simple point-estimate models• Designed techniques to utilize and measure impact of aggregated client data (Gradient’s Core Differentiator)• Leveraged LightGBM and AutoGluon including NLP/Deep Learning on text

Team lead of 2 Data Scientists and 1 Data Engineer• Led hiring of 2 data scientists, designing the interviews and aligning skill-sets with road map needsArchitected dynamic underwriting engine. Framework based on Economic decision theory• Created a road map carving out key areas of ownership and laying out incremental enhancementsOwner of home price prediction model• Spearheaded thought-leading research with University of Chicago Financial Mathematics department• Developed novel description of long-term stochastic home price evolution• Implemented home price diffusion model (Geometric Brownian Motion) from FHFA publications• Cultivated frontier-level expertise in home price indicesDeveloped a new financial product variant from original conception. Delivered beta-ready product underwriting solution in excel using VBAOwner and developer of foundational product performance infrastructure in R• Authored several foundational white papers• Began as self-directed side project, but became existentially critical for capital raising efforts• Established authoritative historical performance & risk analysis for a new asset classCo-owner of portfolio simulation tool developed in Python. Fed by proprietary 180GB housed in SnowflakeKey contributor to credit risk model using XGBoost machine learning algorithmRecommended underwriting criteria policy changes – increased addressable market by 10%Authored Covid-19 investment strategy white paper and performed market ranking analyticsMet with vendors and evaluated value-add of new 3rd party data & analytics productsContributed to call scripts explaining the math of the product, working cross-functionally with head of salesConsulted with external legal counsel to better understand risk of loss in bankruptcy scenariosReviewed investment agreement contract, correcting an error in payment calculation clauses – 0.02% portfolio lift

Led the Data Science initiative for Equator Pro Smart Matching Recommendation Tool• $20M+ value added for mortgagees (1.8% portfolio lift)• Advised product leadership, clarifying product vision and proposed new data driven direction• Responsible for research, prototyping, and delivery of production-ready solution modules• Developed novel algorithm for measuring real estate agent performance• Worked cross-functionally with engineering to implement real-time solutionActed as Project Manager for implementation of FHA & VA loan modification programs• Projected to process $100M+ in loans annually• Assumed responsibilities included: interfacing with client, requirements gathering from government sources, coordinating deployment with engineering team, and leading acceptance testingOwner of Distressed Asset Decisioning Models• Architected modular structure underpinning full suite of collateral-based resolutions: Foreclosure, Short Sale, Deed-in-Lieu, Selling Occupied Properties, Cash for Relocation, and Claims Conveyance without Conveyance of TitleImplemented machine learning models for home price correction and carrying cost estimationDeveloped algorithm to aggregate multiple independent machine learning model predictions together. Improved key metrics by ~10% above simple averaging of model results

• Tutored students of all levels, from elementary school through college, in various areas including Calculus, Statistics, Physics, Geometry, Algebra, Trigonometry, and SAT/ACT test prep.