Hoang Ngo

Developed inference APIs for Google Gemini, accelerating time to first token by 30% through streaming and response pipelining. Improved LLM output consistency across 4 product surfaces and cut compute cost 15% via request batching and caching.

Engineered backend services for Google Pay transaction processing and risk evaluation, processing 3k+ QPS fraud detection pipelines and reducing fraud losses by 8%. Reduced critical payment path latency from 220ms to 160ms and lowered production issues by 30% with rollout controls and safeguards.

Built React and TypeScript tools for healthcare ML systems aligned with Google Health, integrating NLP for medical text analysis and cutting manual review effort by 25%. Developed model evaluation and data validation dashboards, reducing analysis time by 40%, and improved frontend performance for large datasets (35% lower render latency).

Directed architecture and development of HIPAA-compliant healthcare backend services, delivering core systems from 0 to production. Built integration layers and CI/CD pipelines, improving onboarding time by 25% and enabling scalable deployments.

Delivered full-stack enterprise applications across finance and retail using TypeScript, Java, Python, and React. Improved key endpoint request latency by 20% and query performance up to 30% through backend optimization and relational database schema tuning.

Earned a Bachelor of Science in Computer Science at the University of Greenwich from 2011 to 2014.