Anmol Vijaywargiya

A multidimensional, knapsack-based promotions engine algorithm that helps compute the best offers to give customers based on the items in their carts. It handles all sorts of complex offers and helps the users provide the best experience to their customers.

The solution used initially had a complexity O(n^m), and the algorithm I developed got it down to O(n^3). This meant that a calculation that took three minutes (or, at times, ran infinitely) would now take less than five seconds. This algorithm was eventually rolled out successfully for all the major retailers in the country.

A Python and BigQuery-based ETL application to transform a retailer's data into a format that our databases understood. This application was used in scenarios wherein the retailer didn't have incremental updates regarding the product catalog and promotions data changes and, instead, sent in a complete dump of the data every day via FTP. If done traditionally, this meant crunching around 10 million records and replacing them on our databases each day, which was time-consuming and not cost-effective.

Leveraging the ability of BigQuery to crunch gigabytes of data in a matter of seconds, I developed a solution that calculated a difference between yesterday's and today's data before transforming it and pushing it onto our databases. The process, which otherwise would have taken almost half a day to complete, was now accomplished in about 30 minutes.

I developed a dynamic, recursive cataloging system that allows for highly customizable menus with hierarchical components for each menu item. For example, in the case of a pizza, customers can choose from various crust types, each with multiple options for toppings, add-ons, sauce, and cheese.

This structure can be thought of as a tree, with each item having multiple options and each option having multiple items. The system is designed to enable the addition of new menu items, categorize them with labels, specify time slots for their availability, and manage inventory levels.

As a sole developer, I created the dashboard for restaurant administrators and the backend to support the system. The platform was successfully rolled out to various cafeterias and restaurants, leading to its widespread adoption by the organization.

A Ruby on Rails-based solution for the recreation of GCE VMs for about 250+ applications under our team purview in Gojek. I solo-architected and implemented the solution, which comprises the following APIs:
• Recreate infrastructure.
• Switch traffic between new and old infrastructure sets.
• Delete infrastructure.

This solution was used for the migration of VMs from Ubuntu 16.04 to Ubuntu 20.04, since Ubuntu 16.04 had reached its LTE support deadline.

I created an identity management portal using React, MobX-State-Tree, and Ruby on Rails to provide product teams with a centralized location for accessing all aspects of the engineering platform. This user-friendly interface enables users and services to manage their entire lifecycle, from discovery and onboarding to setup and migrations.

While React was a new technology for me, I took on the challenge as we had no front-end developers on our team. I quickly familiarized myself with the platform and delivered an MVP within two weeks. In addition, I helped set up the infrastructure for the front-end and back-end services using Google IAP, Google Load Balancer, and Nginx. For authentication purposes, we added Google SSO.

The portal received positive feedback from the organization, and I found the project to be an enjoyable experience.

I served as the CTO and co-founder of a B2B FMCG marketplace in India, where my primary responsibility was to design and implement a platform that enabled seamless operations and facilitated the scaling of niche FMCG product sellers across state borders. The platform had two key components: a Spring Boot-based backend built on event-driven architecture and a React and Typescript-based SPA that provided a user-friendly interface for customers and operations to interact with the backend.

To tackle the common challenges that arise when managing an FMCG distribution business in India, the platform managed various processes such as the order lifecycle, purchase order lifecycle, inventory, users, payment collections, and on-ground deposits. Additionally, we incorporated a finance module to help the finance team with reconciliation. Through a collaborative approach involving the customers and operations, I developed both aspects of the product and ensured that it scaled effectively as per the needs of the business.

I built Ziggurat, a framework built to simplify Stream processing on Kafka. It can be used to create a full-fledged Clojure app that reads and processes messages from Kafka. Ziggurat intends to abstract out reading messages from Kafka, retrying failed messages, and setting up an HTTP server.
All the teams in Gojek use this library to set up and run Clojure-based apps.