AI Multi-Modal Part Recognition Tool

I need a proof of concept for an AI multi-modal Fusion AI Part Search tool. This tool should leverage various AI techniques such as visual search, text search, and code scanning to help technicians quickly identify the right parts.

Key Features:
- Semantic search, fuzzy logic, and natural language processing to handle variations, abbreviations, or typos.
- Image recognition: Users can take pictures of parts, and the AI will identify them.

Compatibility:
- Must work on mobile devices, desktop computers, and web browsers.

Tech Stack:
- Open to any AI frameworks or libraries (TensorFlow, PyTorch, etc.)

Part Types:
- Must recognize mechanical parts, electronic components, and boiler parts.

Ideal Skills and Experience:
- Strong background in AI and machine learning.
- Experience with image recognition and semantic search technologies.
- Familiarity with developing cross-platform applications.
- Knowledge of NLP and fuzzy logic algorithms.

Looking forward to your innovative solutions!

I have attached a solution picture from another software provider do give you a idea of what I am after