AI/ML Engineer for Generative AI Structural Variation POC

I am looking for an experienced AI/ML engineer to develop a small Proof of Concept (POC) for a private technical project. The core objective is to test whether a generative AI system can perform controlled structural variation of a reference image. The system should take: 1. A base geometric reference image. 2. A natural-language text instruction. It should then generate multiple structurally different variations while preserving the main geometry and identity of the original reference. The key requirement is STRUCTURAL variation, not simply changing colors, textures, materials or adding visual details. For example, the system should be capable of understanding instructions that modify the structure, proportions, sections, shapes or construction of the reference while maintaining its fundamental geometric relationship. I am particularly interested in engineers with strong experience in: - Generative AI - Computer Vision - PyTorch - FLUX / Stable Diffusion - ControlNet - IP-Adapter - Image-to-image pipelines - ComfyUI - Hugging Face Diffusers - Python I already have a technical specification document and reference examples that explain the desired behavior in more detail. The first stage is ONLY a technical POC. The objective is to determine whether the core AI capability works reliably before developing a complete interface or production system. I would like the POC to demonstrate several outputs from the same reference using different structural instructions. The initial POC budget is approximately $200–250 USD, depending on the proposed technical approach and scope. If the POC successfully demonstrates the required capability, there is potential for a larger follow-up project to develop the complete system. Before starting, please briefly explain how you would approach this technically and which models/workflows you would consider for the POC. This is a private technical project. Detailed specifications and reference materials can be shared after initial discussion.