Devon Current

Built React/Redux dashboards and warehouse visualization tools driven by real-time telemetry via WebSockets. Led authentication consolidation across three internal applications, implementing shared login flows and standardizing authentication behavior. Migrated a legacy service from Python 2 to Python 3 and aligned backend data flow patterns with Redux-based frontend state management. Developed and debugged features across JavaScript/TypeScript frontends and Python services using Docker environments and ELK-based logging. Managed QA testing and streamlined pull request processes for a 7-engineer UI team working across three major repositories, improving code integration efficiency.

Building operational dashboards, visualization tools, and configuration interfaces for warehouse automation systems. Contributed to real-time data visualization, authentication standardization across multiple applications, and modernization of legacy services. Built React/Redux dashboards and warehouse visualization tools driven by real-time telemetry via Web Sockets. Led authentication consolidation across three internal apps, implementing shared login flows and standardizing authentication behavior. Migrated a legacy service from Python 2 to Python 3 and aligned backend data flow patterns with Redux-based frontend state management. Developed and debugged features across JavaScript/TypeScript frontends and Python services using Docker environments and ELK-based logging. Coordinated cross-repository integration, QA testing, and pull requests while collaborating with a 7-engineer UI team.

Full-stack developer in a fast-paced field office (20-30 engineers), delivering software solutions for national-security research projects spanning AI/ML, simulation, data visualization, and system optimization. Key developer on a $2.5M project, leading design, React front-end, and back-end optimization. Built RL environments and ML pipelines using OpenAI Gym, Stable Baselines, and automated experimentation tooling. Designed and implemented a Bayesian optimization system that reduced RL model tuning time by 80%. Architected a petabyte-scale time-series data platform integrating PostgreSQL, Neo4j, and real-time telemetry pipelines supporting ML experimentation. Directed a $130k research effort supporting reinforcement learning experimentation in a customized simulation environment.