Statistician by training. Builder by habit.
I’ve spent 14 years inside organizations that needed practical results more than perfect architecture: metrics leadership could trust, models that moved revenue, and clear calls on what was worth building next.
I’ve been the first data scientist at a fast-growing infrastructure company, run the data science and analytics function at a venture-backed real estate company, built credit models for consumer lending marketplaces, and today lead data science at a YC-backed AI agent platform. Along the way I’ve used AI and machine learning not as buzzwords but as working tools, most recently to evaluate AI voice agents and to ship internal tools in a fraction of the time.
Through my firm, Assumed Normal LLC, I bring that same experience to small and medium-sized businesses. Most already have the data they need. They need someone senior to make it reliable, turn it into decisions, and show them where AI genuinely helps.
- Based in
- Austin, TX · working with clients remotely and locally
- Education
- M.S. Statistical Practice, B.S. Mathematics/Statistics, Carnegie Mellon
- Toolkit
- Python, SQL, dbt, PySpark/Databricks, BigQuery, R, Go, scikit-learn, XGBoost, Claude Code
- Domains
- Lending & credit risk, real estate, sports, infrastructure, AI agents
Career
Fourteen years across startups, public companies, and government. Full resume available on request.
Veritus Agent
Head of Data Science
- Lead data science for a next-best-action system combining machine learning with experimentation across the consumer lending lifecycle.
- Analyze AI voice agents to find performance gaps and improve conversation quality, compliance, and recovery rates.
- Use Claude Code and related AI tools extensively to build internal tools that make the whole team faster.
Upstart
Staff Data Analyst
- Credit risk modeling and analytics: probability of default, loss given default, recovery, and segment-level risk measurement.
- Built and monitored production credit models in Python, PySpark/Databricks, and SQL/dbt.
Divvy Homes
Principal Data Scientist · Head of Data Science & Analytics
- Led the function: technical direction, project planning, technical reviews, and regular cadences with non-technical stakeholders.
- Built a rent estimator end to end that became the core pricing signal across operations, and a home value model with an operator feedback loop.
- Shipped a classifier that decides when an AI-based property management platform needs a human response.
- Maintained hundreds of data models for marketing, operations, and competitor tracking.
Supergood
Data Science Advisor
- General data science advisory, including anomaly detection.
Marcus by Goldman Sachs
Associate Data Scientist
- Owned the direct-mail acquisition framework: response models, conditional risk models, and return optimization for every campaign drop.
- Defended targeting and models to the firm’s risk managers; set measurement and diagnostic standards.
- Managed acquisitions through Credit Karma’s Lightbox product in a six-month partnership.
MLB Advanced Media
Senior Data Scientist
- Consulted to MLB teams on ticket buyer profiles, customer revenue health, and re-purchase prediction.
- Forecast revenue for candidate schedules; built umpire call-accuracy models adopted by the Statcast group.
Cloudflare
Data Scientist · first data science hire
- Stood up the function from scratch; built Go services and Dataflow/Spark pipelines feeding Kafka, Pub/Sub, Elasticsearch, and BigQuery.
- Delivered executive analyses and automated revenue tracking across billing and accounting systems.
TransUnion
Senior Statistical Analyst
- Consumer risk models and cross-portfolio attribute leveling across regions.
U.S. Census Bureau
Mathematical Statistician
- Variance estimation, sample validation, and data editing for the Survey of Business Owners.
How I work
A monthly advisory retainer for prioritization, review, and senior judgment, with hands-on projects scoped separately. That keeps incentives tied to outcomes, not billable hours.
I’m direct about what will and won’t work, I write things down, and I leave behind systems your team can run without me.
Outside of work I live in Austin with my wife and two dogs, and I still scrape and analyze odd datasets for fun.
Want to talk it through?
I’m happy to look at what you have and tell you where I’d start.