Reclaimed 8+ PB annually by architecting an SQS-backed storage lifecycle platform that automatically hides and later safely deletes unreferenced S3 media, eliminating months of manual engineering oversight per year

Led delivery of a low-latency live video ingest and distribution pipeline using ECS, built to support professional sports leagues and serve 1.3M SRT/RTMP streams annually

Migrated critical flows of media microservice processing 50K multi-hour videos daily to modern .NET on ECS

Shape cross-team technical direction through design reviews, POCs and partnerships with AWS and Dolby

Delivered video splitting and trimming backend, saving Support ~600 hours per year and releasing to users soon

Mentor engineers through onboarding, pairing, code review and one-on-ones

Architected autoscaling and queues for migration of encoding infrastructure—which ingests hundreds of hours of video per minute—to Terraform-managed EC2, saving $450K annually and eliminating a single point of failure

Served as Ops Manager, refining Hudl’s incident response processes, expanding observability and leading retros

Built core components of an event-driven video hour calculation service to enforce storage limits for 100K teams

Scaled media metadata MongoDB cluster to support seasonal traffic spikes by sharding and adding read replicas

Reduced video buffering for millions of users by building a custom MP4 preloader