Capital One
Richmond, VASenior Java Full Stack Developer
AI agents & automation · Payment platforms · Application security
Built payment fulfillment services, production AI tooling, and reusable agent workflows in a regulated banking environment.
Role details & technologies +
- Developed a Python/FastAPI LLM inference service that summarized historical payment-file processing, surfaced patterns and anomalies, and supported end-of-day operational decisions through a Slack bot.
- Maintained an internal AI skills registry and CLI for Claude Code and Windsurf; authored reusable Jira, pull-request, Confluence, and CI/CD skills adopted by partner teams.
- Applied agentic workflows to code generation, refactoring, and test scaffolding; shared reusable prompts and skill patterns across engineering teams.
- Built the public-facing Fulfillment API (Java 17, Spring Boot 3.x on ECS Fargate) secured with OAuth2, mutual TLS, and field-level encryption for sensitive payment data.
- Led the first production migration to Java 21 / Spring Boot 4, hardened with Resilience4j circuit breakers and SQS retry/fallback queues.
- Built vendor file ingestion on AWS Step Functions (40+ vendor formats) and the EMR Spark pipeline generating TSYS posting files with automated reconciliation.
- Contributed to a real-time money movement platform in Go — event-driven, outbox-patterned, active/active across two AWS regions.
- Engineered data pipelines with Databricks PySpark, Flink SQL, and Kafka, publishing governed datasets to Snowflake and Microsoft OneLake.
- Remediated Checkmarx, SonarQube, and Mend findings and dependency CVEs; implemented least-privilege IAM, Secrets Manager/Vault, and PGP encryption aligned with PCI DSS and NIST 800-53 controls.
Environment: Java 17/21, Spring Boot 3.x/4.x, Go, Python, TypeScript, React, Apache Spark (EMR), Databricks, Flink, Kafka, AWS (ECS Fargate, Lambda, Step Functions, DynamoDB, S3, SQS, EventBridge, EMR, Glue), Snowflake, OneLake, Docker, Jenkins, Splunk, OpenTelemetry