AI/Computer Vision Developer for Car Customization App

We are building a car customization app and need someone who can help us solve a specific problem. Users upload six photos of their car: front, rear, both sides, a 45-degree front view, and a 45-degree rear view from the opposite side. They can then customize the car with spoilers, bumpers, wheels, and wraps. The challenge is keeping these changes accurate across all six images. Our current Gemini-based approach produces misplaced parts and modifications that look different from one angle to another. We need the selected parts to fit correctly and look consistent across every view while preserving the original car’s shape and details. Asking users to mark incorrect areas and keep regenerating is not enough. We would like to start with a proof of concept to test your approach before moving into the full app. Please share any relevant work involving multi-view image editing, computer vision, or 3D reconstruction. In your proposal, briefly explain how you would approach this problem, what inputs you would need, and your estimated timeline and cost for the proof of concept. Be clear about any limitations. Technical questions: 1- How would you make sure the same spoiler, wheels, or bumper appears in the correct position, size, and perspective across all six photos? Would you use image editing, a shared 3D model, or a combination, and why? 2- How would you test whether the customizations are consistent across all views and the original car’s details remain unchanged? What could cause your approach to fail, and how would you handle those cases?