Andrew Long

Led data-driven experiments and model improvements that increased model efficiency by 30% and prediction accuracy by 25%, and drove production integrations that avoided deployment performance degradation.

Owned frontend platform and Python/Node services to raise SLO attainment and ship safely, reducing MTTR 45% quarter-over-quarter and cutting rollback frequency by 60% while improving mobile time-to-interactive by 22%. Coached a team of 6–8 engineers on on-call, design docs, and release practices.

Built FastAPI services and React surfaces to support ranking and catalog reads, maintaining p95 under 100ms at peak and reducing refund tickets ~70% via idempotency and outbox/inbox patterns. Consolidated REST calls into GraphQL, improving time-to-interactive by 22%.

Built NLP and LLM-based systems to convert resources and auto-grade reading fluency, increasing content creation speed by 40% and cutting manual grading time by 60% while enabling real-time feedback at scale.

Developed experiments UI, CI a11y gates, and Kubernetes services to improve experimentation velocity and storefront reliability; reduced onboarding time, decreased schema drift bugs ~40%, and cut MTTR 35%.

Architected real-time bidding optimization and model-serving infrastructure with Go, PyTorch, ONNX, and Kubernetes, increasing advertiser ROAS 18% and reducing infrastructure cost 28% while sustaining sub-100ms bid response SLAs.

Developed AI-powered engines and ML pipelines that improved targeting ROI by 20% and prediction speed by 35%, optimized backend processing to reduce times by 40%, and lowered storage costs by 25% with AWS integrations.

Built scalable ML systems and feature stores for omnichannel bidding, improving targeting lift 9%, reducing compute overhead 20%, and cutting incidents 22% via enhanced monitoring and autoscaling.

Developed core bidding and prediction systems in C# and Go and automated ETL/validation pipelines, improving bid accuracy 7% and reducing training data errors 25%.

Improved backend reliability and performance using C#/.NET and C++, reducing API error rates 32% and p95 latency 27% while accelerating root-cause analysis 38% through enhanced telemetry.

Improved test automation and CI/CD to reduce manual testing and deployment times by ~70%, designed cloud-native microservices for Xbox Live improving processing time by 30%, and built diagnostic tooling reducing issue resolution by 40%.

Built scalable backend services and telemetry pipelines on Azure, improving reporting latency 30% and reducing incident resolution time 20% through real-time processing and optimized schemas.

Designed automated API validation suites and data validation scripts integrated into CI, reducing post-release defects 25% and shortening release cycles 20%.

Built ML pipelines for genre classification and metadata generation that cut manual tagging by 50% and improved metadata consistency by 30%, and deployed FastAPI services and dashboards to automate metadata submission and quality assessment.

Completed a Bachelor of Science in Computer Science focused on software engineering and machine learning fundamentals.

Completed an Associate of Science in Computer Science preparing for transfer to a four-year CS program with coursework in programming and systems.