Ohm Patel

I led the creation of a scalable Ads delivery pipeline that supports dynamic product Ads and complex‬ product listings, enabling seamless integration of shopping features into the Reddit Ads delivery‬ pipeline.

I architected this system to handle massive data loads efficiently, using parallel processing,‬ system-level memory optimization, and high-performing data structures to manage and serve millions of‬ product ads with minimal latency. My contributions directly increased product ad impressions to around 10 million per‬ day and helped secure new advertisers, leading to 30,000 USD of revenue per day. I demonstrated‭ extraordinary leadership in optimizing and scaling our Ads Pipeline. Under his guidance, the pipeline was‭ re-architected to handle up to 25 million products efficiently. This scaling effort involved complex‭ optimizations to the data processing and storage frameworks, allowing us to reduce memory footprint and‭ save costs by 50% compared to the old implementations while supporting a substantial increase in‬ product inventory. This work has directly contributed to an increase in ad revenue and improved‬ click-through rates.‬

A project that leverages an interplanetary file system to store files in a decentralized network and track and maintain the files' metadata on the Ethereum blockchain. It uses ganache and Go-Ethereum for development purposes.

Pay Later is a POC application for leading fintech company Simpl (https://getsimpl.com/), enabling users to make immediate transactions with partnered merchants. Users can pay later (monthly) based on their billing. The idea behind this project was to model the architecture, designing the database models and microservices. The project was entirely built using Golang.

A Twitter-like back-end app built using Golang, MySQL, and GraphQL for querying. The app can log in/signup. A user can post a Twitter-like two-sentence text post, follow another user, and view their timeline. The app is deployed using Docker Compose into multiple containers of the GraphQL server and database server.

I built the entire custom integration of McDonalds Restaurant's Menu and Order System worldwide with‬‭ Europe's biggest food delivery company, Deliveroo.

I helped build one of the most advanced, scalable‬‭ , and real-time order processing systems for any food delivery company in the world out there for one of‬‭ the biggest restaurant chains worldwide, McDonalds. The system accepts orders placed on the Deliveroo‬‭ app, transforms the order payload into the Deliveroo system, stores the information, and directly injects the‬‭ order into the POS system of the McDonald's Restaurant. The complex system involves implementing‬‭ various technologies such as Kafka and Snowflake for real-time processing of orders and terabytes of‬‭ scaling. This system handled around 95,000 per day on average, with the highest from the UK market‬‭ being 44,000 orders per day. The system built has a success rate/up time of 99.03%, which has a‬‭ direct impact on Deliveroo's revenue. This entire system handles 10% of the entire‬‭ volume of orders daily for Deliveroo. I designed and implemented a scalable and robust system that has reduced our order‬ failure or incidents by 20%.