Scott Crawford

I’m a staff-level engineer with 12+ years at Google building distributed systems at YouTube and payments scale. I help platform teams ship reliable, high-throughput infrastructure faster—from architecture through live observability—with hands-on ownership across Java, Go, and full-stack delivery.

At YouTube, I owned end-to-end design and delivery of YouTube’s account recovery system, rebuilding the Java/gRPC backend and React frontend to cut average recovery time by 40% while handling 10M+ monthly requests. I also scaled user data protection infrastructure on Bigtable to support GDPR and CCPA compliance across 2B+ active accounts without degrading read/write performance.

I’ve repeatedly modernized systems for performance and operability: migrating monolithic account management to gRPC microservices on GKE reduced p99 latency by 35% and enabled independent team deployments. I instrumented services with OpenTelemetry, built Prometheus dashboards with Cloud Logging alerts, and established CI/CD standards with canary rollout policies—cutting post-deploy incidents by roughly half.

More recently, I integrated Vertex AI embeddings and vector search into account recovery flows, reducing related support ticket volume by 18%, and built Pub/Sub-driven pipelines with Dataflow for faster detection of compromised account patterns. I use AI tooling like Claude Code and Cursor daily to compress the design-to-deploy cycle across complex codebases.