Building infrastructure for data at scale.

Tintri builds a highly performant and flexible data storage product that plugs into customer's infrastructure to provide insights into applications workloads. Like a public cloud service, the product simplifies application provisioning without the user struggling to understand the underlying storage details.Supported the team that built the scalable and highly performance filesystem, storage platform and hardware. Delivered multiple product families including the very first all flash appliance.

Data Domain builds a high performance enterprise data protection appliance with industry leading active data reduction technologies. My roles included- Technical lead for roadmap and delivery high-end multi-controller Global Deduplication Array.- Technical lead and implementation of data and metadata layer for variable sized content defined boundaries for efficient deduplication.- Performance improvement for linear scaling of nodes.- Source side deduplication to improve write performance and scaling.- Asynchronous data replication for multi-controller systems.

NetApp builds an industry leading data storage appliance to simplify cloud and enterprise data management needs. I was part of the core filesystem team (WAFL) with the following roles- Technical lead for data de-duplication and compression in the WAFL file system- Technical lead for data containerization for storage efficiency- Development engineer for remote caching volumes - FlexCache

Developed the Honeycomb product, a scale-out storage product aimed at large scale archival market. The system featured a novel integrated and searchable metadata that allows easy organization and access to large repositories of unstructured data. My contributions were in the following areas:- Product definition and features- Design and implementation of an object based file system and access protocol- Integrated metadata indexing and search- Cluster management and storage platform

Early engineer in Zambeel's team to build a highly scalable, fault-tolerant and scale-out Networked Attached Storage (NAS) service. My roles included developing- Multi-node, fault tolerant layer offering a durable data service- Distributed volume and management layer to simplify system operation- Storage resource management system to track and utilize hardware for fault tolerance

Engineer in Microsoft's Hotmail scalable data backend.Service - Built high performance and scalable incoming e-mail serving tier using a lightweight data flow design instead of traditional multi-threaded approach. This maximized hardware utilization and lowered overall cost.Performance and Efficiency - Improved performance of the backend cluster with better queuing, locking and synchronization, detecting idle or aborted connections. This increased throughput by 25% and greatly reduced cluster sizes needed for peak traffic.Reliability - Improved availability of service by active load balancing of incoming traffic. This improved the success rate of incoming emails by over 50%

Implemented a scalable gateway for high bandwidth data transfer between IBM mainframe clusters and new high-speed commercial printers. The goal was to manage and process a very high inflow of data and match it with the hardware capabilities.Built SCSI tape driver for proprietary tape hardware. Designed and implemented a new tape file system for efficient backup and restore.High performance file system for data spooling and sharing of data between different system architectures.