Krylov - ML Platform: Part of the team building and managing highly scalable, distributed in-house AI/ML model training platforms - Krylov with at-par features and all bells and whistles of a modern ML platforms such as Google’s AI platform or Microsoft’s Azure.

⏺ Features for platform robustness: DAG execution of tasks, workflow features, configurable retries, configurable compute capacity, easy to access logs, metrics

⏺ State management using apache zookeeper managed by the team

⏺ Connectivity to Hadoop platform and features store

⏺ Custom workload and platform metric creation and collection using Kubernetes libraries and Nvidia agents (GPU metrics)

⏺ Training/experiment metric collection and dashboarding using prometheus and grafana

Near Real Time data processing and data pipelines: Design and build highly efficient Near Real Time (NRT) data pipelines towards building feature stores and big data for computer vision features such as visual search, vector similarity search and user uploaded content moderation (video, image and text )

⏺ Distributed Event driven systems built using Apache Kafka based off of eBay platform for data streaming

⏺ Built to scale: Systems scaled to produce and process billions of NRT everts per day (scale of eBay listings and text/image/ video uploads) with negligible processing lag and optional compute capacity.

⏺ Efficiency: Employed data aggregation techniques (Flink) and non-blocking reactive programming to achieve high efficiency and processing speed

BigData Platform: Build and maintained eBay's in-house Hadoop platform. eBay has world class hadoop clusters with more than 3000 nodes and petabytes of data. Features include Spark , Hive, HBase, data governance tools, data intelligence tools, job scheduling and tracking tools etc.

Data platform: Part of the team that maintains and enriches eBay's Data access layer(DAL). A highly available system that manages eBay's database infrastrure and resposible for site availablility.

⏺ Leading a team building Java based systems for the retail inventory management, Supply chain warehouse management, logistics, invoicing and billing

⏺ Responsible for design, development, deployment to production and production performance and application monitoring

⏺ Full stack development experience - from provisioning VMs/ writing Chef cookbooks to design development and deployment of application in TDD and pair programing environment

⏺ Test Driven Development with AGILE Application Development scrum team with pair programming

Part of the success story to migrate Califoria's EDD from legacy systems to modern web system with zero disruption to customer experience

Completion of health care reform related changes for 2 of the leading health insurance providers in the US

⏺ Leading efforts for End to End solution from Requirement Gathering, solution design and implementation.

⏺ Design and development of new web based systems as well as integration of newly developed systems with the legacy systems and creation of interfaces

⏺ Experience working in AGILE and Waterfall methodologies

⏺ Implementation of web MVC, Web services, SQL Data bases, Spring services, SOA, JSP, JSDL and Java script

⏺ Experience in working with all phases of SDLC

⏺ Team leading and Joint Application Design experience

⏺ Experience working with geographically distributed teams including offshore teams