Dec 2020 – Sep 2021 · 9 mos · Cupertino, California, United States

Jan 2016 – Dec 2020 · 4 yrs 11 mos

• As a Platform Team member, I define, design, and implement new algorithms, architectures, and frameworks for various mission-critical services running in Amazon Cloud (AWS). I am also involved in troubleshooting and resolving critical issues to ensure availability. I help teams gain insights into their services by identifying bottlenecks and providing solutions to circumvent them. Part of the Data Platform senior leadership building block, object &, file storage System; distributed high-performance key-value datastores, and data integration platforms.
• Developed a scalable and reliable Global Data Storage System that was responsible for object, block, and media-specific storage systems for Netflix Streaming and Studios. This is instrumental in building the next-generation media-aware storage and transport system for the Netflix studios named Global Storage at an Exabyte scale.
• Partnered with Memcached Open Source community and drove various features in it like native encryption, external storage support on SSD/NVMe, LRU Crawler, Memcached Proxy, etc.
• Enhanced EVCache with features like persistence, encryption, cache warming, hashing of keys, auto compression & scaling, data coherency, and eventual consistency.
• Developed resiliency framework so that we can keep serving customers even when most of the services are not available or not reachable i.e. degraded service instead of no service.
• Part of the resiliency efforts to make the cross-region movement of traffic. We are now able to move traffic from one AWS region and rebalance it across other AWS regions in a few minutes.
• Developed systems for security services for credit card fraud detection, account take over, certificate management, etc.

Nov 2011 – Jan 2016 · 4 yrs 2 mos

• Developed and Managed the High-Performance Data System based called EVCache with features like replication(region), zone (datacenter) awareness, encryption, locality, persistence, indexing of data, cache warming, backups, stats collection, monitoring, and reporting.
• Developed resiliency framework so that we can keep serving customers even when most of the services are not available or not reachable i.e. degraded service instead of no service
• Implemented REST request invocation pattern across various services so we can provide a consistent view of the data as a request travels across various services.
• Implemented a request orchestration framework that loads data required to serve a request by calling various services in parallel, thus cutting down the overall response time.

Jun 2008 – Oct 2011 · 3 yrs 4 mos

• I was responsible for the research and implementation of the Netflix home page and various other pages for both the website as well as devices.
• I also worked on the core ranking & scoring algorithms that make personalized movie recommendations. Worked on rearchitecting the monolith app into microservices.
• Was pivotal in introducing and evangelizing Hadoop/Hive at Netflix which eventually evolved to a data platform team focused exclusively on Big Data
• Significantly improved the performances and overall availability of the systems.
• Improved various algorithms that were responsible for selecting movies for the users, this ensured we could handle additional traffic without increasing the capacity.

May 2003 – Jun 2008 · 5 yrs 1 mo

• SigmaQuest is the leader in scalable solutions for Product Quality Management. Our solution, SigmaSure, leverages a powerful root-cause analysis engine to improve product quality and customer satisfaction.
• Unlike home-grown systems and MES/ERP systems, SigmaSure is hardware and software agnostic. This allows SigmaQuest to automatically collect consistent data across the extended enterprise, and perform detailed analysis on terabytes of test, quality & logistics data.
• SigmaQuest customers use SigmaSure to collect data directly from their ODM's/CM's to monitor shop floor activity, alerted to anomalies before they become problems.

Jun 1999 – Sep 2003 · 4 yrs 3 mos

• NextSet's Private Trading eXchange (PTX) is a J2EE based web trading application that enables banks and other financial institutions to offer online trading of financial instruments to their corporate and institutional clients.

Mar 1997 – May 1999 · 2 yrs 2 mos · Kolkata, West Bengal, India