Designed and built a federated Next.js platform integrating 13 bioinformatics agents into a unified workflow interface. Engineered an SSE streaming pipeline with token reconciliation for real-time LLM workflows and seamless multi-day session persistence. Developed a modular SDK + component architecture (OpenAPI, Shadcn/UI, MDX), reducing frontend feature delivery time by 25%.

Optimized rendering pipelines for complex visualizations (50k+ data points), achieving sub-second UI responsiveness and increasing researcher productivity by 35%. Adapted and modularized the FinnGen PheWeb UI to support client-specific data models, redesigning queries and visualization logic while reducing dashboard latency by 20%. Architected a DuckDB-backed Genome Browser by modernizing open-source gnomAD logic, significantly improving high-volume sequence exploration performance.

Built high-performance genomic data exploration tools enabling real-time visualization of multi-million row datasets. Developed Burden Reports using React + Hasura (GraphQL) with server-side pagination and structured querying for scalable analytics. Improved rare-variant discovery workflows by optimizing data accessibility for thousands of researchers.

Migrated corporate infrastructure from JAMStack to WordPress, reducing content update turnaround time by 50%. Engineered containerized Docker deployments and improved performance benchmarks to A+ (GTmetrix).