Marcel Oliveira Alves

An augmented reality iOS app developed during my master's program to help neurosurgeons during preoperative planning. I wrote the app in Swift and used frameworks such as ARKit, SceneKit, Metal, UIKit, Accelerate, Vision, and XCTest.

With ARNeuro, the user can inform the 3D coordinates of three to five points obtained from a computed tomography volume of the patient's head and a target point or region. The app uses ARKit capabilities to find the correspondent points on a real patient. It then performs some linear algebra operations to match the virtual points and the real points, a process known as registration. Finally, the app renders the target point or region overlaid onto the patient's head on the device screen. This enables the surgeon to perform minimally invasive surgery.

My tests indicated a low error rate, around three millimeters from the target. I've also conducted performance tests using Instruments and Xcode. I'm currently improving user experience and app accuracy by testing different renderers, such as SceneKit and Metal.

A project of one of the most prominent players in the entertainment industry. The project contained multiple streaming apps for tvOS, iOS, and iPadOS. Considering all apps, the average rating was 4.6 out of 5.0 from a total of more than 135,000 user evaluations. I was responsible for implementing accessibility features, fixing bugs, and creating new screens and UI components for the project's apps.

The highlight of the project was fixing and documenting a tricky bug that harmed the user experience. The bug happened while the app was running on the Apple TV device but not on the simulator, making it harder to debug. I also opened a bug report with Apple because a bug caused the issue in their AVFoundation framework.

An iOS app for one of the biggest tech companies. The native app, written in Swift, worked as a front end that communicated with the server through a REST API to control a router device and configure user profiles, permissions, login, and more. I was in charge of implementing new features using the MVVM architecture, creating animations, doing code reviews, suggesting changes, and fixing bugs.

Project highlights:
- I fixed a tricky memory leak bug that was happening in some Coordinator classes.
- I suggested a refactor to change the architecture of a specific part of the app from MVVM to VIPER, significantly reducing the time to implement new features for that part of the app, the amount of repeated code, and the number of bugs caused by the view model that complicated state management.

Built the front end of a chat feature for a social network startup iOS app. I implemented one-to-one and group chats using REST APIs, socket programming, and AWS framework for five weeks. Also, I identified and reported bugs on the back end.

A simple Swift app that enables users to record audios, apply effects to their recordings, and share with friends. I mainly developed the app to learn about the Lottie library, AVFoundation framework, and the current policies to publish an app in the App Store. Besides the mentioned technologies, I also used UIKit, Auto Layout, storyboard, and CocoaPods.

An app to help the players of the card-game Force of Will (FoW) by allowing them to search for cards and using in-game tools, such as life counter, dice, or timer. I developed the iOS version in Swift, and my colleagues developed the Android version in Java.

My responsibilities on this project included:
- Writing a Python script to extract the card information from the web to create our database.
- Creating the cloud database using Firebase.
- Designing our local and cloud database approach to synchronize and update.
- Taking product decisions to improve the user experience.
- Developing the iOS version of the app.

The app had to be discontinued from the App Store due to our lack of time for maintaining it.