Microsoft2020 — present
Microsoft Defender
Security services, and moving them onto Kubernetes.
Xbox platform
- Owned core infrastructure for Xbox's shared microservices platform — 116 services on ~45 production AKS clusters (~22K pods), up to ~1M requests/second globally, with no platform-attributable availability incidents across major game launches. Primary owner of the cluster fleet behind 17+ live services.
- Shipped defense-in-depth ingress on Azure (DDoS Protection L3/L4; App Gateway WAF + Bot Manager + custom L7 rate limits; NGINX mTLS/routing); WAF rate-limit rules contained a known L7 DDoS attack and preserved backend stability.
- Hardened the platform for title launches: validated 50,000 RPS per cluster and 3× traffic bursts, onboarded the highest-traffic service to 10× its prior request volume, and resolved a launch-day Sev-1 with custom App Gateway rate limiting.
- Led design and early implementation of the Gateway API and Istio migration: authored the spec and ran the cross-team design review. The review showed edge mTLS terminates upstream, so chose that over sidecar passthrough and set rollout, rollback and scaling decisions, plus a draft PR.
- Made KEDA the platform's default autoscaler through a shared Helm chart and CI/CD pipelines, preserving standard resource scaling while enabling queue-depth and request-rate scaling for custom-metric workloads.
- Owned the observability stack — Grafana migrations, access and upgrades, Managed Prometheus, and the FluentBit log pipeline — and architected its migration into a hardened internal tenant.
- Led the "Protect the Admin" security overhaul: replaced standing Azure access with just-in-time PIM elevation and RBAC across 15+ subscriptions for 50+ engineers, adopted by five sister teams. Migrated 30+ secrets and service connections to managed and federated identity with no downtime, eliminating certificate-based auth.
- As on-call incident owner, cut recurring alert noise and made the remaining alerts actionable; in one cycle eliminated ~30 of 94 recurring alerts by fixing their triggers.
- Ran the platform's cost and capacity analyses, identifying WAF rightsizing and Kubernetes decommissioning opportunities.
- Prototyped a read-only, human-gated on-call DRI copilot that triages incidents against runbooks and returns confidence-scored verdicts backed by an evidence trail. A live incident surfaced multi-region clustering and query gaps that shaped the next iteration.
Xbox Cloud Gaming
- Built C#/.NET session-management and provisioning microservices for console streaming and shipped Remote Play to GA. Integration tests, reconnect handling and telemetry cut reconnect failures ~15% and held provisioning success above 90%.
- Built an internal logging developer platform for the service fleet: a C# ILogger framework with PII redaction and enforced data contracts, adopted across all services. Re-architected the pipeline from fluentd to FluentBit over ~878 pods, reclaiming ~219.5 CPU cores.
- Co-owned readiness for the multi-region core migration across Americas, Europe and Asia-Pacific, covering logging, data-pipeline and integration-test readiness, and took on more of it as teammates rotated off the project.
- Mentored Explore-program interns and ran the team's engineering AMA series: 6 sessions, 361+ attendees, rated 4.6/5.