Best Fit
Staff / Senior Data Scientist
Applied AI, production ML, identity/fraud risk, data quality, and analytics products.
Staff Data Scientist
Current role at Abodemine
Senior Data Scientist
Previously at Prove
PhD Astrophysics
Computational modeling at UC Santa Barbara
AI Data Integration
LLM-assisted schema matching and canonicalization
At a Glance
Senior data scientist for AI-heavy data products.
I am strongest in roles where messy real-world data, ML systems, and product judgment all matter. I bring a PhD research background, production ML experience, and hands-on AI system design.
Current Work
LLM schema matching
Mapping client loan tape workbooks to a 1,000+ attribute canonical property and loan dataset.
Proof Points
Models with product impact
99% AL=3 accuracy, 98% fraud caught at 3% friction, and 200% faster model deployment cycles.
What I’m Good At
Building data science systems that survive contact with reality.
My best work happens where models, data pipelines, business rules, monitoring, and human review have to fit together into something reliable.
Applied AI and LLM workflows
Prompting, structured outputs, semantic validation, deterministic rule engines, and human-in-the-loop feedback.
Production machine learning
Model development, deployment patterns, evaluation pipelines, monitoring, drift detection, and maintainability.
Data quality and canonicalization
Schema matching, entity resolution, enum mapping, normalization, and turning inconsistent inputs into trusted data.
Data investigation and insight extraction
Exploring unfamiliar datasets, finding product-relevant signal, identifying limitations, and tracing anomalies back to their source.
Fraud and identity signals
Risk scoring, onboarding signal aggregation, trust models, verification systems, and friction-aware product tradeoffs.
Python engineering for data teams
Reusable repositories, pytest, CI/CD, service patterns, and tools that reduce repeated scripting and fragile workflows.
Scientific thinking
Simulation, uncertainty, validation, technical communication, and disciplined reasoning from computational astrophysics.
Experience
Applied data science with a systems bent.
My work sits at the intersection of modeling, software engineering, and product judgment: building tools that are accurate, observable, and useful to the people who depend on them.
Professional Work
Staff Data Scientist and ML practitioner
Currently Staff Data Scientist at Abodemine, where my work includes matching client workbook schemas to an internal canonical schema with AI and LLM-assisted workflows. Previously, I grew from Data Analyst to Senior Data Scientist at Prove, where I worked on trust, identity, model monitoring, international scoring, and production analytics workflows.
Read more about my experienceCurrent AI Work
Using LLMs to make messy client data usable.
At Abodemine, I work on AI-assisted schema matching: mapping client workbook fields into an internal canonical schema so heterogeneous data can move through consistent downstream systems.
Problem Space
Schema matching for real-world workbooks
Client workbooks rarely arrive with clean, predictable field names. I help design systems that infer intent from column names, workbook context, examples, and business rules.
AI Application
LLM-assisted canonicalization
The work combines prompt design, structured outputs, validation logic, and reviewable confidence signals so AI suggestions can fit into production data workflows.
Why It Matters
Practical AI for data operations
This is the kind of AI work that matters in industry: reducing manual mapping burden, improving consistency, and making complex ingestion workflows easier to scale.
Older personal projects are still available in the project archive.
Research Foundation
Computational astrophysics, translated into industry data science.
I earned my PhD at UC Santa Barbara studying white dwarf binaries with theoretical and computational models. That training still shapes how I approach uncertainty, simulation, validation, and technical communication.
Explore publications and researchGet in Touch