iOS Developer (SwiftUI / Vision API) – Build a Real-Time Movement Tracking Camera (1-Week Sprint)

We are looking for an experienced native iOS developer to build a high-performance, reusable camera UI component. The module needs to capture real-time video, process human body landmarks using Apple’s native Vision framework, and overlay dynamic UI elements on the user’s joints in real time without dropping frame rates. This is a strictly time-boxed, one-week sprint. The final deliverable must be a clean, well-documented Swift package or modular component that can be easily dropped into our existing broader application architecture. Scope of Work & Deliverables This project is broken down into a tight 7-day timeline with specific expected deliverables: Days 1-2 (Architecture & Camera Setup): Implement a custom camera preview using AVFoundation and seamlessly integrate it with a SwiftUI view hierarchy. Ensure zero memory leaks during session start and stop cycles. Days 3-4 (Vision Integration): Implement the VNDetectHumanBodyPoseRequest API to capture and parse human body joint coordinates in real-time. Days 5-6 (UI Overlay & Polish): Build a fluid SwiftUI overlay that maps the Vision coordinates to the screen. Draw smooth, lag-free lines and nodes connecting the detected landmarks. Day 7 (Testing & Handoff): Refactor for performance to maintain 60 frames per second, ensure compatibility with iOS 16+, and provide a standard README document with implementation instructions. Required Tech Stack Language: Swift 5.0+ UI Framework: SwiftUI Native Frameworks: AVFoundation, Vision API Architecture: MVVM or a similar clean architecture pattern Ideal Candidate Profile Proven track record of building performant, native iOS components. Deep understanding of memory management when dealing with live video feeds and CoreMedia buffers. Ability to write clean, maintainable, and highly documented code. Clear daily communication regarding sprint blockers and progress. Screening Questions Please include brief answers to these questions in your proposal to help us evaluate your technical fit: Describe your approach to passing real-time video buffer data from AVCaptureSession to the Vision framework without bottlenecking the main thread. Can you share a link to a previous project or GitHub repository where you built a custom camera interface or worked with Apple's native Vision APIs? How do you typically handle SwiftUI state updates for high-frequency data to avoid unnecessary view redraws?