Jul 2018 – Aug 2024 · 6 yrs 1 mo

• Designed and deployed Tier-4 data center hardware infrastructure for General Purpose Cloud, AI/ML HPC GPU Cloud, and Archival Storage, supporting rack densities from 5kW to 42kW.
• Enhanced IaaS reliability by implementing L11 server acceptance testing, component-level fault detection and prediction systems, and zero-downtime maintenance orchestration using live VM migration, automated upgrade workflows, and rack-awareness constructs.
• Designed and built custom server hardware for Flipkart across multiple generations of Intel and AMD CPUs and NVIDIA GPUs, fostering deep technical collaborations with server, component, and firmware manufacturers to optimize performance and efficiency.
• Built an end-to-end cost analysis pipeline to measure and optimize utilization efficiency across private cloud services by developing accurate TCO models, unit cost derivations, metering systems for hardware allocation, and integration with utilization metrics.
• Transformed Flipkart’s storage infrastructure from converged systems to composable disaggregated architectures, improving cost efficiency, availability, and scalability. Developed an in-house software-defined storage system using external SAS JBOD/Fs, scaling it to 500+ PB, later transitioning high-performance workloads to NVMeoF-based network-attached storage.
• Designed and developed an in-house highly available, low-latency, and cost-effective network-attached software-defined storage solution leveraging SPDK, NVMeoF/RDMA, and dual-ported SAS/NVMe drives.
• Publications & Conferences
• Slash N 2022 - “Building a Cost Effective and Highly Reliable Private Cloud Infrastructure”
• Cloud Native Data Management Day NA 2021 - “Building Disaggregated Storage Infrastructure”
• OCP Tech Day India @ Flipkart 2019 - “Flipkart's Infrastructure Journey”
• IEEE CLOUD 2018 - “Migrating VM Workloads to Containers - Issues and Challenges”

Jan 2016 – Jun 2018 · 2 yrs 5 mos

• Designed and developed a comprehensive hardware lifecycle management system for an in-house IaaS stack, incorporating Zero Touch Provisioning and automated maintenance orchestration.
• Optimized virtualization infrastructure to spin up VMs in under 20 seconds using disk snapshots and lightweight images.
• Built hardware observability tools leveraging SES, SMART, EDAC, Mcelog, BMC SEL, and SDR for enhanced monitoring and diagnostics.
• Developed a normalized REST API proxy for IPMI 2.0, scaling it to efficiently manage BMC operations across 20k+ heterogeneous fleet of servers.
• Engineered a high-performance job scheduler for conducting service health checks on 50k+ virtual machines.

Aug 2012 – Dec 2015 · 3 yrs 4 mos

• Systems and Reliability Engineer for end to end infrastructure stack including hardware, operating systems, virtualization, databases and applications.
• Designed in-house and implemented in-house data center infrastructure management system (DCIM) for a heterogeneous compute, storage and networking devices.

Aug 2008 – Jul 2012 · 3 yrs 11 mos · Bangalore

• Systems and infrastructure engineering for Yahoo! Application Platforms, Yahoo! Developer Network, Yahoo! Mobile Platforms, Yahoo! Maps and Geo Location Platforms.