Senior Data Scientist and Software Engineer

Led development of behavioral risk models for suspicious account and transaction activity using Python, SQL, scikit-learn, anomaly detection, graph analytics, clustering, and statistical analysis. Designed evaluation methods for highly imbalanced datasets with incomplete labels, investigated false positives and false negatives, and improved detector precision while preserving emerging risk signals.

Built near-real-time feature pipelines from high-volume event streams with Kafka, Spark, PostgreSQL, and AWS. Productionized machine learning services using Docker, Kubernetes, Terraform, CI/CD, CloudWatch, dashboards, and automated validation. Mentored engineers and data scientists, documented model methodologies and tradeoffs, and evaluated OpenAI, GPT, Claude, RAG, and agentic systems for analyst-assistance workflows.

Analyzed large-scale service telemetry using Python, SQL, statistical analysis, and time-series methods to distinguish anomalies from expected infrastructure variation. Built anomaly-detection pipelines across distributed metrics and event streams, applying clustering and robust baselines where incident labels were incomplete.

Developed production data workflows with Python, Spark, Kafka, distributed systems, CI/CD, containers, Kubernetes, and GCP-style cloud infrastructure. Performed detector error analysis, created monitoring strategies for alert quality and behavioral drift, and contributed reusable analytical libraries, design reviews, and observability standards.

Developed Python and SQL analytics for cloud-service telemetry, using statistical modeling and anomaly detection to identify unusual resource, authentication, and workload behavior. Built time-series features and classification workflows with scikit-learn, pandas, and distributed data pipelines, and designed offline evaluation datasets for noisy, rare, and incompletely labeled operational events.

Integrated analytical methods into production services using Azure, Docker, Kubernetes, Terraform, CI/CD, automated testing, and deployment safeguards. Investigated detector errors through feature-level analysis, log correlation, and controlled experiments, while documenting assumptions, experiments, and monitoring criteria for engineering teams.

Built Python and SQL analytical software for large scientific datasets, applying statistical analysis, clustering, time-series techniques, exploratory data analysis, anomaly detection, and signal processing. Developed workflows for sparse and noisy measurements with rare positive events and incomplete ground-truth labels.

Created maintainable data-processing services using Linux, Python, C++, batch pipelines, automated testing, and later containerized deployments. Supported production research systems through debugging, performance profiling, CI/CD automation, infrastructure scripting, and reproducible technical documentation.

Bachelor of Science in Computer Science.