Work
Current work, selected initiatives, writing, talks, and career. Skip to now, featured work, initiatives, side projects, publications, speaking, or career.
Current role
Streamlining incident and change management for GPU cloud operations. Establishing on-call tooling, change templates, and the Grafana → PagerDuty → Jira alert path. Also setting up Major Incident guidelines so teams can respond with clarity when the stakes are high.
Featured work
A $100M engagement with 15 Tbps of required bandwidth, delivered in three months. The work involved a new network architecture, capacity across the storage stack, and load validation from customer GPUs.
$100M · 15 Tbps · 3 monthsSelected initiatives
Grew SRE and Platform Engineering for OCI Object Storage from 5 to 45 across four geographies. Shifted the operating model so service teams consumed platforms instead of building their own.
5 → 45 engineers · 4 geographiesProduction rollbacks dropped 60% per service. Build times dropped 80%. Change-caused incidents dropped 50%. The leverage compounds when service teams stop reinventing region builds, pre-prod qual, and fleet management.
60% / 80% / 50% reductionsBuilt AI-driven support and automation workflows across the incident lifecycle: detection, triage, remediation. Reduced ticket volume by 30% and improved response times by 50%.
30% fewer tickets · 50% faster responsesBuilt the repeatable region-build framework that enabled OCI expansion from 4 to 100+ regions, with consistent reliability and operational readiness at each launch.
4 → 100+ regionsAutomated patching and vulnerability management for the 50,000+ node fleet, achieving 100% compliance with FedRAMP standards without slowing engineering velocity.
50K+ nodes · 100% FedRAMPSide projects
Live cloud GPU pricing across 10 providers (AWS, GCP, Azure, Lambda, RunPod, Vast.ai, CoreWeave, Paperspace, Hyperstack, Salad). A Claude-powered advisor picks the cheapest reliable rig for your workload, with push alerts when prices drop.
A simple way to send a restaurant meal to a friend or pay one forward for someone in the community. Diners can also request their favorite restaurant and help bring it onto the platform.
AI × SRE, every day, on the mic. Short daily takes on outages, tools, and AI agents in production.
Publications
Daily notes and shorter posts on the Writing page.
Speaking
Career
Streamlining incident and change management for GPU cloud operations. Establishing on-call tooling (PagerDuty, Jira, change templates), wiring the Grafana → PagerDuty → Jira alert path, and setting up Major Incident guidelines.
Led the global SRE organization for OCI Object Storage — 45+ engineers and 3 managers across the US, India, UK, and Mexico, supporting 100+ regions and a 50,000+ node fleet. Owned end-to-end service reliability, incident lifecycle, automation strategy, and operational scalability for a business that grew to $600M ARR, 10 EB of customer data, and 100% YoY growth.
Led release and operational readiness for Object Storage across multi-region deployments. Built the region-build framework that enabled OCI expansion from 4 to 100+ regions, and developed automation for fleet patching across 30,000+ nodes.
Twenty-one years delivering enterprise implementations across regulated industries. Consistent on-time delivery with measurable defect reduction.