Caleb Oneyemi

A modular monolithic Node.js API for creating and managing a mock premier league. Admins can create, edit, and delete teams and fixtures, while regular users can log in and view teams and fixtures.

API supports auth, caching, rate limiting, and a bunch of other common operations. I built the project from scratch as a simple template to structure back-end Node.js projects.

A web-based application built out for internal revenue operations and asset retrieval teams across four countries to handle day to day workflows.

In a rapidly moving team with tons of complex real-world operational challenges, I had to focus on improving the flexibility and reliability of the system to address those real-world issues. There were a couple of major issues we had when I joined, and these included service degradation/downtime and failed operations from external systems we integrated with.

Over time, I played a role in addressing some of these challenges and more. We initially used Grafana/Loki for monitoring, and given that some of our microservices were making synchronous calls to other services, this was not enough. I collaborated with my technical lead and the DevOps team to introduce Prometheus for metrics and Jaeger for tracing to help debug and resolve degradation issues. I also introduced queuing and retry mechanisms to resolve the occasional failed operations from external systems.

Besides these, I worked on other feature additions along with moving the contract management component to a dedicated service out of the core revenue operations back-end service.

A back-office CRM for financial advisors with the goal of unifying workflows and providing a simple all-in-one toolbox, thereby saving cost and time, reducing complexity from switching between multiple tools, and overall improving the user experience.

With this goal in mind, I joined the team and built out three microservices in TypeScript with a focus on integrations over a period of about four months. I also improved data import performance and reduced CI pipeline costs, which were two critical issues when I joined.

I helped lead the migration of older services to TypeScript since I was the first engineer on the team with previous TypeScript experience, and I also replaced our internal RabbitMQ library with an open-source AMQP wrapper. These actions helped simplify configuration and enabled us to spot errors earlier, thereby increasing the overall reliability of the system.

I joined primarily to build out the social media component, which initially just handled posts. Some work had been done on media uploads, but it was highly coupled to posts.

I improved the handling of posts, decoupled the existing media upload implementation from posts, and extended it to handle uploads for other Lambda-based services I implemented including clips, comments, and moments (like Instagram stories). I also stabilized the overall API by fixing errors and adding tests, which took coverage from a little under 20% to over 90%.

A Web API that fetches the latest transaction data from Ethereum and pushes it to all listeners via Socket.IO.

The API allows subscribers to configure the events they listen to. They can listen to all transactions or transactions related to a specific ETH address.

The tech stack used includes Node.js, TypeScript, TypeORM, PostgreSQL, Socket.IO, Express, TypeDI (for dependency injection), Docker, and Docker Compose.