I am part of the tech-infra team, that's creating Software Defined Infra and Services in our new DC. I used own and drive the tech mandate for the End-to-End experience team. I've designed, built and owned multiple platforms and systems that are widely used at Flipkart to deliver critical functionality. Some of them include...
IaaS Compute: I own and am responsible for Flipkart's Elastic Compute Platform. This along with our ELB are fundamental IaaS primitives on top of which All our PaaS components and customer applications are deployed. Our recent work has centered around improving utilization through better fitment and to address the tech/domain debt accumulated in our run up to last year's Big Billion Days.
Scotty: Deployment and change management service that enables an immutable infra paradigm in the cloud. It coordinates and governs infra and release assets to achieve, correct, predictable, well controlled deployments and change management on SDI(Software Defined Infra). It is designed to do it's job without communicating with atomic units of infra which makes it trivially scalable and able to handle change management for 100s of thousands of infra assets, simultaneously.
Repo-svc: A generic post build release service that's used to publish and consume binary release assets in a well defined manner. Release artifacts are published into versioned repos which can further be used to compose versioned environments which make up the Universe of released assets for associated deployments. The design of repo-svc allows for repeatable and predictable deployments by enforcing well defined immutable environments comprised of release assets.
Lego: Lego is a highly performant, highly scalable (in terms of traffic and developers) and resilient application framework. It is an incredibly flexible framework that is used by multiple teams at Flipkart and is deployed in different shapes on over a thousand bare-metal servers. These installations are combinedly capable of handling upto 2 billion requests per hour. This is one of the critical systems that make the festive season and events like Big Billion Day possible at an unprecedented scale.
Igor: Is a platform to curate upto a 100,000 online stores comprised of millions of products while processing billions of product updates. It delivers low latency entity processing with tremendous visibility and fine grained control over the data processing pipeline. It provides strict processing and fairness guarantees in a multi-tenant setting while minimizing operational overhead. All of Flipkart's stores are managed by this platform.
Drove gamification and Social Network Analysis research and solutions in the lab. Developed a service delivery automation solution. Co-authored an FP7 joint proposal for smarter government initiative involving social networks. Authored multiple patents and publications
Co-founded Spinaxys with two friends after taking a sabbatical from IIIT Hyderabad. We provided high IP services to organizations for our cash flow and built products on the side. I was responsible for product conceptualization, design and development. Our products included a collaborative filtering service for pushing relevant sms content, world's first open-source digital signage and blue-casting system and an enterprise knowledge management solution.
I setup the company's technical team and operations. Built the tech and QA teams, setup development, build and test infrastructure, implemented an auditable project management process system for the Dev and QA team. I acted as the technical architect and presales solution architect. We reached profitability in 6 months with a 4 fold increase in the number of employees.
Published four papers in top tier computer vision conferences and a dissertation; Mentored and guided students at the lab. My research was centered on Content Based Image Retrieval with focus on building highly scalable CBIR systems. My main contributions to the field are part of my dissertation 'Efficient Image Retrieval Methods For Large Scale Dynamic Image Databases'. I was Involved in organizing an international computer vision conference (ACCV 2006).
Suman Karthik, Efficient Image Retrieval Methods For Large Scale Dynamic Image Databases, Dissertation, MS by Research, CVIT IIIT Hyderabad, September 2010
Chandrika Pulla, Suman Karthik and C.V. Jawahar Efficient Semantic Indexing for Image Retrieval Proceedings of 20th International Conference on Pattern Recognition (ICPR'10),23-26 Aug. 2010, Istanbul, Turkey.
Suman Karthik, C.V. Jawahar, Incremental On-line semantic Indexing for Image Re- trieval in Dynamic. Databases 4th International Workshop on Semantic Learning and Applications, CVPR, 2008, Florida
Suman Karthik, C.V. Jawahar, Analysis of Relevance Feedback in Content Based Image Retrieval, Proceedings of the 9th International Conference on Control, Au- tomation, Robotics and Vision (ICARCV), 2006, Singapore.
Suman Karthik, C.V. Jawahar, Virtual Textual Representation for Efficient Image Retrieval. Proceedings of the 3rd International Conference on Visual Information Engineering(VIE), 26-28 September 2006 in Bangalore, India.
Suman Karthik, C.V. Jawahar, Efficient Region Based Indexing and Retrieval for Images with Elastic Bucket Tries, Proceedings of the International Conference on Pattern Recognition(ICPR), 2006
Targeted Maintenance of Computing Devices in Information Technology Infrastructure United States US 13/551,015
System And Method For On-Demand Simulation Based Learning For Automation Framework United States US 13/604,846