•Built and scaled core platform features using TypeScript and Next.js to support reinforcement learning environments and outcome verifiable task systems
•Developed internal tools to source, vet, and onboard domain experts efficiently for large scale human data projects in advanced science and other high value domains
•Implemented dynamic dashboards and workflow systems to manage operationally complex AI data pipelines
•Optimized server side rendering and application performance to support data intensive, research focused workloads
•Integrated APIs and backend services to power benchmarking systems used by frontier AI labs
•Contributed to internal tooling that supported the development of high difficulty reasoning benchmarks evaluated against frontier models
•Collaborated cross functionally with researchers, operators, and leadership to ship features in fast paced, high ownership environments
•Participated in architectural decisions to ensure scalability, reliability, and long term maintainability of the platform.