Mehmet Efe Akça

A very extensible open-source Rust program to organize files arriving in a folder. Its primary use case is to organize files by their download URLs but can be extended with custom actions and matchers written in JavaScript. It can remember files that were downloaded before via hashing. It can also show the file after organizing or conditionally show a prompt for a user to choose a file's new location.

This project was a personal project to make my life easier by organizing downloaded files. Building this made my coursework folder much cleaner as it could automatically delete redundant files and organize without my input.

The first time I built this program, it was very restrictive and couldn't customize behavior without changing its source code. After a round of refactors, I integrated a JavaScript engine into the program and implemented a new architecture with generic matchers and actions that apply to individual files. With this new architecture, the program combined the reliability of a Rust-based codebase with the extensibility and freedom of JavaScript-based tooling.

The project was about an Electron application with linear video editing tools made for creators. I was a front-end developer building both the UI for the application and the video editing logic. Some of my significant contributions included:

• Optimizing silence cutting logic to improve performance by over ten times in pure JavaScript.
• Building a clip placement algorithm for a non-linear editing UI and writing a performant two-dimensional collision algorithm in an HTML division.
• Building a custom video player layer on top of VideoContext to show a timeline preview with minimal performance penalties in JavaScript.

A React, Firebase, and TypeScript full-stack platform with a React Native mobile application, brand panel, and an admin panel, reaching 100 thousand users at its peak. I was the lead and only developer for most of the project's lifetime, from conception to production. I coordinated with product owners and built multiple flows and features, including:

• Influencer registration and OAuth 2 integrations with Twitter, Instagram, and Tiktok.
• Brand advertisement creation flow, allowing brands to choose categories and influencer filters.
• Admin flows for verifying influencers, managing advertisement campaigns, and verifying advertisement entries.
• Influencer flows for applying to campaigns and managing an application to a campaign.

While building this, I worked with a NoSQL database, Firestore, and had several technical challenges in balancing cost and performance. I had to ensure that influencer filters could be performed with the least cost by reducing the number of fetched documents from the DB and with the best performance by creating performant composite indexes. The monthly cost of the project was around $15 at its peak.

An Electron-based app for day trading that combines multiple data feeds and allows a trader to make better decisions.

I started as a performance engineer and then became a software architect in the team. I built fundamental architectures such as the multi-window system, Redux-based state management system, and indicators system and aided in developing many other features. This allowed the app to become multi-window, persistent, and performant.

I also used Rust to build a prototype application for a market replay product that processes and caches market data and allows for seeking the data stream. This project involved using Async Rust and the challenges of building a performant trading application.

A React Native application that helps volunteers feed street animals (namely, dogs and cats). It's a mobile application that allows its users to define feeding routes, points, and more and helps them keep track of when and how much feeding needs to be done. It was built in an effort to help local communities that feed street animals.