Infrastructure and database reliability engineer with 14+ years operating critical systems at scale. At Block (Square), owned the Tier 0 Items service (a 256-node, 50TB+ MySQL/Aurora cluster serving every Square merchant globally) and led 5 major zero-downtime infrastructure migrations over 5.5 years. Earlier, designed and operated Vitess on Kubernetes at Stitch Labs before Vitess became PlanetScale and delivered a public conference talk on the architecture. Personally operates AS400719, a multi-site BGP network peering at 4 internet exchanges, managed as infrastructure-as-code with Ansible.

• Owned the Items service, Square's mission-critical catalog system backed by a 256-node, 50TB+ MySQL/Aurora cluster across 64 active shards, serving every merchant globally
• Led 5 major zero-downtime infrastructure migrations: on-prem to AWS on Kubernetes, Active/Active multi-region reads, Elasticsearch to OpenSearch (16M docs in 1.5 hrs), 50TB+ database resharding, and RDS to Aurora
• Designed and led development of AuthProxy, an internal Go/Kubernetes credential management platform enabling 6 teams to integrate 12 third-party APIs (DoorDash, Uber Eats, TikTok Shopping, Meta, Google, etc.) without building OAuth flows; manages 1M+ credentials with 500K automated refreshes/hour at 50 req/sec sustained
• Architected Active/Active multi-region read path across US-East and US-West, delivering 70ms latency reduction for east-coast merchants and unblocking the Tier 0 Orders team's own multi-region migration
• Served as Terraform SME across 6+ service teams; owned IaC for each team's full stack (Aurora, Redis, OpenSearch, Datadog, PagerDuty, Fastly CDN), kept configurations current as the platform org rolled out infrastructure changes, and trained 15+ engineers to eliminate single-point-of-failure dependency on myself
• Built the oncall onboarding program and runbook for the Channels team from scratch; designed shadowing schedules that onboarded 20 engineers, with the program later adopted by other teams across Block
• Led oncall response for 20+ high-value merchant escalations; diagnosed and resolved an ElastiCache CPU credit exhaustion SEV, migrated to predictable instance types, and introduced monitors that prevented recurrence

• Migrated 3TB sharded MySQL database to a 97-shard Vitess cluster running across 300+ Kubernetes containers, one of the earliest production Vitess-on-Kubernetes deployments; delivered Percona Live 2017 talk on the architecture alongside Vitess's creator from YouTube, cited by the Linux Foundation when Vitess joined the CNCF
• Reduced infrastructure spend by ~$50K/month within 2 months with no reduction in availability or application performance
• Managed 100+ virtual hosts via Ansible; implemented GCP Cloud Trace for distributed tracing across 15 microservices; unified authentication across entire infrastructure
• Transitioned on-call from ops-only to full engineering team ownership; led team of 5 DevOps engineers

• Grew engineering team from 6 to 18; owned hiring across frontend, backend, and DevOps; implemented Scrum achieving 90%+ on-time delivery QoQ
• Returned to hands-on infrastructure work as Director of Platform Engineering

• Operate a multi-site BGP network (AS400719) with 4 IPv4 /24s and 3 IPv6 /48s; peer at 4 internet exchanges (FCIX 10G, FREMIX 10G, ARIX 1G, Edgoo 10G) plus Hurricane Electric transit
• Manage 8 VyOS routers (6 edge, 2 VRRP HA core with OSPFv3 backbone) and 30+ Linux hosts across colocation and home sites, entirely as IaC via Ansible
• Colocation stack: Proxmox VE clusters, Kanidm identity management, LibreNMS network monitoring, Akvorado flow analysis, HashiCorp Vault, InfluxDB