Senior Infrastructure Engineer · AI Systems & Reliability
I operate at the intersection of large-scale distributed infrastructure and AI system reliability — managing 1,200+ hypervisors, 24,000+ VMs, and Kubernetes clusters that keep production AI workloads running. I'm fascinated by where AI systems silently break and what it takes to build ones that don't.
I'm interested in how large-scale AI infrastructure constraints shape what AI systems can actually do in production — specifically, the gap between what models are theoretically capable of and what they reliably deliver under real operational conditions.
My work managing distributed compute — Kubernetes, OpenStack, 1,200+ hypervisors, 24,000+ VMs — and responding to production incidents gives me a ground-level view of where AI systems break, degrade, and fail silently.
I want to bring that operational depth into research on AI system reliability, observability, and agentic system robustness. I believe the most important work in AI right now isn't capability — it's making capable systems trustworthy at scale.
B.E. Electronics & Telecommunication Engineering
North Maharashtra University
After managing K8s clusters running AI inference workloads, I've catalogued the failure modes that don't throw errors — they just degrade. Here's what I found and how to detect them early.
Operating one of the region's largest VMware environments taught me that scale doesn't just multiply problems — it creates entirely new failure categories. A post-mortem of what I learned.
A structured approach to root cause analysis that goes beyond the immediate trigger — tracing cascading dependencies, documenting systemic fixes, and building an institutional memory of how systems fail.