AI/LLM Fine-Tuning Engineer – Arabic, Egyptian Arabic, Gulf Arabic & English Smart Home Model

Freelancer Project Message — ~1000 characters Hello, I’m looking for an experienced LLM/NLP Fine-Tuning Engineer to develop a specialized language model for MEGA Smart Home AI. The model should understand natural smart-home commands in: Egyptian Arabic Gulf Arabic English User sentences will be Arabic-only or English-only; Arabic-English code-switching is not required. The main goal is not general conversation, but Natural Language Understanding for smart-home control. The model must identify intent, device, location, action, and parameters, then generate a strict structured JSON output compatible with our MEGA OS. It must also handle ambiguous requests by asking for clarification instead of guessing, and recognize unknown/unsupported devices or commands. Required experience: SFT, LoRA/QLoRA, LLM fine-tuning, Arabic NLP/NLU, Egyptian/Gulf Arabic, dataset creation/annotation, Python, evaluation and benchmarking. Please provide: 1. Previous Arabic/Egyptian/Gulf LLM projects. 2. Recommended base model and why. 3. Fine-tuning methodology. 4. Dataset strategy and expected size. 5. Evaluation methodology and metrics. 6. Final trained model and inference/deployment files. This is intended for local/offline deployment on MEGA hardware, so model size and quantization efficiency are important.