May 2024 – Sep 2025 · 1 yr 4 mos · Mountain View, CA · Hybrid

• Inventory Management System
• Working on Designing and implementing a highly scalable, distributed, and concurrent event based inventory management systems which support operations such as inventory inbound, outbound and picking using Command Query Responsibility Segregation (CQRS) architectural pattern implemented using Java Akka framework.
• Impact: Implement CQRS based micro-service architecture to support 100 million transaction per day for inventory management system.
• Technical Details: AKKA framework allows to process parallel request without obtaining a lock on data by using messaging queue to serve incoming requests. Java based backend system uses distributed Redis cache, Dynamo Db, apache Kafka queues.

Feb 2021 – Apr 2024 · 3 yrs 2 mos · Seattle, Washington, United States · On-site

• 1. Responsible for optimization of product details page list view on Echo Show devices for better shopping
• 1.1 This project helped Alexa Shopping team provide a new scalable shopping experience to customers in US marketplace and replaced older experience by providing comprehensive product detail page.
• 1.2 Implemented a java based solution which optimized memory usage by updating necessary
• information for in-focus items on screen using sliding window approach when a user navigate on candidate list.
• 2. Responsible for redesigning Alexa Shopping search and buy offer presentation service involving multiple teams to reduce new features development, testing and deployment time.
• 2.1 Redesigned presentation service that facilitated Alexa shopping and it's partners deploy, test and debug changes much faster than legacy system without sacrificing quality of software. Deployment time reduced from 2 weeks to few hours.
• 2.2 Designed a new configuration based architecture which was feature agnostic allowing partner and developer to build features at an accelerated pace.
• 2.3 Designed a micro-services architecture in java to break down a legacy monolithic search and buy offer presentation service.
• 3. Responsible for world wide expansion of multi-turn voice prompt assisted shopping experience for Echo Dot devices
• 3.1 Expanded multi-turn voice prompt experience from US marketplace to global markets. This project help reduce customer abandonment rate by more than 70%
• 3.2 Implemented features for users to follow up shopping queries with commands like "Add to cart", "buy now", "next or previous item". This greatly improved previous experience which just added first result to cart without conformation or follow ups.
• 3.3 Implemented service to asynchronously send shopping related notification Alexa app for screen-less devices using amazon SQS.

Apr 2017 – Jul 2019 · 2 yrs 3 mos · Bengaluru Area, India

• Cross Dock for Grocery Delivery :-
• Details: I worked on developing a distributed grocery delivery supply chain system which supported multiple user-preferred delivery time throughout the day.
• 1. In this model, large line haul vehicles deliver orders from warehouse to smaller vans near delivery addresses and vans are dispatched for delivery to destination address.
• 2. This project included various teams like shipping, logistics, order management system, payment system. Each team implemented micro-services such as order management, route planning, return/replacement order etc to support this new supply chain delivery system.
• 3. Implemented micro-services which provided updates on orders status from order management system to downstream system like logistics and shipping.
• 4. Distributed supply chain systems communicated using messaging queues with other systems to provide live order updates to facilitate ground operations.
• Call Bridge System :-
• Details: I worked on designing and implementing a call bridge service which connects customers with delivery persons without sharing their phone number to enforce privacy of customer data.
• 1. Designed and created a database to support this system using mysql. Implementation was done in Java using spring boot framework
• 2. Partitions of records older than 6 months were dropped from Database using cron jobs which ran during night hours.
• 3. Ingested each new record entry asynchronously at time of creation in a H-Base storage which was used for analytics purposes and fetching information for records older than 6 months.
• 4. Master database records were duplicated into two slave replicas and most of the read operations were performed from these slave databases. A master database standby was also used in case of failure of master database.
• 5. Mapping for around 0.5 to 1 million customer numbers were generated daily and the system handles a load of around 3 million calls per day.

Jan 2016 – Dec 2016 · 11 mos · Bangalore

• Software Engineer
• URL Shortener and Targeting: -(Project link - http://zop.link/)
• Details: Developed a URL shortening service which provides short aliases for redirection of long URL with additional features like URL re-targeting based on OS, device type and region, deep-linking with fallback URL, QR Code and UTM Parameter analysis.
• Marketing Campaign
• Details: Development a system to send broadcast and personal emails and push notifications for marketing purpose. Cron was setup to schedule jobs for sending push notifications to mobile application. System send 0.1 million notification in 15 mins.