Mostafa Mohamed Emad

I developed a real-time machine learning (ML) model serving application focused on online ranking for both chains and individual users. This service maintains a recommendation retrieval latency of under 250ms and handles a throughput of 10,000 requests per minute (RPM). The application effectively serves around 1 million users daily, providing nearly 3,000 chain rankings per request, ensuring efficient and timely recommendations.

A fraud detection rule engine system.

The system is divided into two parts: a rule engine and a rule repository. Rules are developed separately and pushed into the rule engine using a cron job. The rule engine receives orders using Kafka topics. Every order is passed into the fraud detection process, along with running all rules to verify whether the order is fraud or not. Then, the status is reported back.

An online banking application. The application aims to provide online banking features for clients associated with FAB bank.

Some of the features included are:
• Sign up for an account, credit card, or loan.
• Manage accounts, debit, and credit cards.
• Integration with multiple third-party resources.

Mazadat is an eCommerce app for buying and selling safely, introducing bidding features for used items to be sold for the first time in the Egyptian market. I worked on this project as a full-stack developer using Node.js for back-end development and React/Redux for front-end development. We built this app from scratch till production. Later, the app was upgraded into a microservices architecture design to scale up.

What makes this app different is that it adds safety to the process of conducting cash and item exchanges. It isolates the seller from the buyer entirely and, at the same time, adds bidding. Bidding was only to be executed in Egypt on Facebook pages, which added the risk of the seller and buyer meeting up in person to complete a cash and item transfer.