Open to opportunities

Ashton Cothran

@ashtoncothran

Message

I build governed ML systems for credit, fraud, pricing, and real-time financial decisions.

United States

Message

What I'm looking for

I'm looking to lead applied AI and machine learning work that improves financial decisions, reduces risk, and meets rigorous governance standards. I value roles where I can build production systems, mentor engineers, and partner closely with risk, compliance, and product teams.

At Ramp, I own business-critical ML models supporting more than 1 million annual transactions across credit decisioning, fraud detection, and financial intelligence. I improved fraud-detection precision by 18–25% and led experimentation that increased approval quality by 12–15% without increasing portfolio risk.

I build explainable, regulatory-ready systems with monitoring for drift, stability, and bias, reducing post-deployment incidents by more than 30%. I also led engineers in productionizing low-latency cloud-native decisioning systems operating at sub-100ms latency.

Previously at Capital One, I built credit underwriting and fraud models supporting 20M+ customer decisions annually, improved model AUC by 6–10%, and helped deliver regulatory model examinations with zero critical findings. At Strip, my fraud and transaction-ranking work contributed $8–12M in annual fraud-loss reduction and a 4% conversion uplift.

I started in data analytics and data science at JPMorgan Chase, using Python and SQL to support risk reporting, model governance, and high-impact policy decisions. I also bring hands-on GenAI experience, using RAG, embeddings, and agentic workflows to reduce internal analysis turnaround time by 40–50%.

Experience

Work history, roles, and key accomplishments

Current

Principal Applied AI & ML Engineer

Current

Ramp

Jun 2022 - Present (4 years 2 months)

Owned business-critical ML models supporting 1M+ annual transactions across credit decisioning, fraud detection, and financial intelligence. Designed and deployed explainable ML models that improved fraud detection precision by 18-25% and led LLM-driven analytics and agentic workflows.

Machine Learning Fraud Detection Credit Decisioning Model Monitoring SHAP LLM RAG Embeddings AWS Sagemaker Lambda Bedrock Python SQL PyTorch MLFlow Airflow CI CD

Applied ML Lead

Capital One

Sep 2019 - May 2022 (2 years 8 months)

Built and maintained credit underwriting and fraud detection models supporting 20M+ customer decisions per year. Developed feature engineering pipelines and orchestrated batch and near-real-time inference pipelines, improving decision throughput by 3x.

Feature Engineering Model Monitoring Python SQL Spark AWS Airflow CI CD Machine Learning Credit Underwriting Fraud Detection

Senior Machine Learning Engineer

Stripe

Dec 2016 - Aug 2019 (2 years 8 months)

Developed ML models for fraud detection and transaction ranking, driving $8-12M in annual fraud loss reduction. Executed A/B tests validating ML-driven improvements, resulting in 4% conversion uplift.

Machine Learning Fraud Detection Transaction Ranking Feature Engineering A B Testing Python SQL AWS PyTorch scikit learn

Data Analyst/Data Scientist

JPMorgan Chase & Co.

May 2013 - Nov 2016 (3 years 6 months)

Delivered SQL- and Python-based analytics over TB-scale enterprise financial datasets supporting risk reporting and evaluation for 10+ business units. Automated end-to-end analytics workflows, reducing manual reporting effort by 30-40%.

Data Analysis SQL Python Risk Reporting Hypothesis Testing Governance