Jan 2019 – Jan 2022 · 3 yrs · India

• In my role as Lead Engineer in Walmart's Platform team, I've contributed to various telemetry initiatives.
• Re-architecture of the alert notification system
• This is essentially an event driven system where users subscribe to alert events of interest. Subscription mechanism allows not just simple filtering that is part of most streaming frameworks but also advanced patterns such as temporal patterns that is offered using a CEP engine.
• Identified the bottlenecks in old system and designed/built modern solution by splitting into independent, isolated, scalable components, which has coasted through peak holiday loads with no issues, while maintaining high performance/low latency.
• Recommendation Engine
• What started as a hack idea gained interest and became a telemetry solution available in production.
• A smart platform that sends insightful alerts by enriching with contextual changes (recent changes, service dependency info etc) and action recommendations (classifying the issue and thereby matching with recommended resolution actions)
• Built out the generic platform modules and the smart AI core, that uses both classification algorithms and neural networks.
• During this time, I was able to file couple of patents pertaining to messaging (improving robustness of data pipelines by seamless redelivery of missing data) and the recommendation engine.

Feb 2017 – Apr 2018 · 1 yr 2 mos

• Part of the Messaging platform of the Flipkart Cloud Platform (FCP) group
• Contributed to the development of new features as well as making the platform more robust to progress towards goals of maximum availability, reliability and scalability.
• Implementation of constructs to isolate/minimize failures, avoid potential choke points in the message broker, scaling up of database system through sharding, setting up of backup and disaster recovery mechanisms etc.
• Achieved system throughput of over 100K messages/sec resulting in smooth running even
• during Flipkart’s annual Big Billion Days with no downtime and no message loss.

Jan 2016 – Jan 2017 · 1 yr · India

• Goodbox is an m-commerce app that offers SMEs a platform to create custom mini-apps to connect with customers. It enables enterprises to go online in no time in an easy, economically viable fashion and for users, it offers a mega-app with a large set of options across diverse categories.
• Designed & developed from scratch a scalable, modular, extensible, pluggable rule engine system. The intent was to use the system for a wide range of use-cases including automated chat response, marketing communication, custom business logic etc. Declarative mechanism allows product or business analysts to customize behaviour at runtime. Components included Kafka message broker, Camel routing core, MySQL as metadata store, spring expressions based rule interpreter and a pluggable handler module to allow easy plugin of 3rd party written custom handlers into the engine.
• As a member of the core backend team, I contributed to the codebase which form the backbone of the goodbox platform, including core web APIs, batch jobs for user notifications, marketing communication, fulfilment systems etc.
• Designed and developed goodbox wallet solution.

Apr 2015 – Sep 2016 · 1 yr 5 mos · Bengaluru Area, India

• Design & development of search solution for account data using Solr, which provides a plethora of query patterns. Worked on the implementation of a fault-proof indexing pipeline as well as a generic API for search.
• Design/implementation of a new Notifications platform, with the aim of solving typical event/observer pattern, for e.g; notifying interested customers of specific item price change, notifying sellers with demand alerts etc. Asynchronous integration paths included kafka streams, storm layer for stream aggregations/other kinds of processing.
• Contributions to a new smart platform, meant to enhance customer experience with data analytics. Worked with data in HDFS, pipelines for data transport across systems, Cassandra for real time queries, bit of mining algorithms for classification etc. Purpose was to use tremendous amount of data to come up with insights that could be fed to the user-facing systems to enhance a user's online experience.

Jul 2013 – Mar 2015 · 1 yr 8 mos

• Contributed to the Customer Accounts project of Walmart Global eCommerce.
• Mutli-tenant, low latency, large-scale restful web services. Cross DC deployment topology that serves site traffic in the order of dozens of millions on a normal day.
• Implementation of data migration solutions for moving data out of legacy systems to the new platform via various channels, with heavy use of Camel EIP.
• Development of various logging/monitoring/alerting mechanisms to ensure high availability of critical systems using python scripts, mongoDB, ElasticSearch, anuglarJS UI etc

Jan 2013 – Jun 2013 · 5 mos

Jul 2012 – Dec 2012 · 5 mos