Robert Brown

Senior AI Engineer with deep experience building production ML and generative AI systems for financial services. Built LLM and RAG tools for financial document analysis that cut manual review effort by 30% and improved credit risk prediction accuracy by 15%. Ready to apply this background in Python, SQL, AWS, Kubernetes, and MLOps to improve risk decisions and automate financial workflows.

Senior AI Engineer at CAPITAL ONE (2025-02 – Present)

• Built production AI systems with TensorFlow and SQL that accelerated loan approvals and improved credit risk assessments.
• Developed LLM and RAG applications using embeddings and vector search to automate financial document analysis, accelerate insight extraction, and reduce review time.
• Built scalable AI services with Python, AWS, Kubernetes, and MLOps, enabling reliable, near-zero-downtime production deployments.
• Improved risk prediction accuracy by 15% through feature engineering and model optimization.
• Reduced manual review effort by 30% by automating financial analysis workflows.
• Partnered with data scientists and engineers to productionize AI solutions, improving reliability and scalability through model monitoring.

Senior Machine Learning Engineer at CAPITAL ONE (2022-08 – 2025-02)

• Developed and deployed PyTorch ML models for credit risk, lending decisions, and customer analytics, improving risk assessment accuracy.
• Built Python, SQL, Spark, and Snowflake pipelines processing millions of financial records daily, improving data availability for analytics.
• Boosted model performance through feature engineering, rigorous experimentation, and validation using PyTorch and Snowflake.
• Deployed and monitored production models using Docker and CI/CD, achieving 99.9% availability.
• Reduced inference latency by 35% through model and serving optimization.
• Automated model training and retraining with CI/CD, accelerating production updates and improving deployment efficiency.

Machine Learning Engineer at SCALE AI (2018-02 – 2022-08)

• Developed ML and NLP solutions that improved enterprise document-search accuracy and efficiency.
• Built Python, Spark, and Airflow pipelines, cutting data preparation time by 50% and enabling daily model retraining.
• Deployed containerized ML models with Docker and Kubernetes on Azure, delivering scalable, high-availability inference at 10,000+ requests per second.
• Automated model delivery through CI/CD, cutting deployment time from hours to minutes and improving deployment reliability.
• Increased production reliability by 30% and reduced model latency through expanded pytest coverage, Prometheus monitoring, Grafana dashboards, and performance tuning.

Data Scientist at ACTIVTRAK (2015-11 – 2017-12)

• Developed Python and TensorFlow predictive models that increased forecast accuracy and enabled faster, data-driven decisions.
• Built and deployed analytics and ML solutions that accelerated analysis, improved decision-making, and delivered measurable business impact.
• Built scalable Snowflake feature pipelines with Transformers, improving ML data quality and accelerating model development.
• Automated weekly reporting with Airflow and Python, cutting manual work by 80% and improving data quality through standardized validation checks.
• Integrated ML models into Java applications, enabling real-time predictions and improving application capabilities.

Data Analyst at ACTIVTRAK (2013-08 – 2015-11)

• Built and optimized SQL datasets, dashboards, and reports, enabling faster, data-driven business decisions.
• Improved data quality by automating validation and data processing in Azure Data Factory, reducing errors and enabling reliable analytics.
• Prepared and optimized clean, structured Snowflake datasets, accelerating analytics and machine learning workflows.

Bachelor of Science in Computer Science – Rice University (2009 – 2013)