AI/ML Engineer Needed – Debug and Complete Existing AI Agent System (Python, Qwen, OpenClaw, Proxmox

I am looking for an experienced AI/ML engineer to help troubleshoot, validate, and improve an existing AI-powered autonomous task prioritisation and scheduling system. The system is already partially implemented and deployed. I need technical assistance identifying inconsistencies, fixing integration issues, and ensuring the deployed components operate correctly together. Existing Technology Stack Python, scikit-learn and FastAPI/Uvicorn Qwen2.5 language models and GGUF quantisation Unsloth fine-tuning using LoRA/QLoRA llama.cpp inference servers OpenClaw agent orchestration Proxmox LXC containers (Debian) Gitea repository integration Scope of Work Review the existing Gitea repository and deployed architecture. Investigate inconsistencies between the fine-tuning notebooks, model metadata and deployed GGUF files. Troubleshoot the unavailable base-model inference backend. Verify correct model loading, routing and API responses. Validate the machine-learning prediction service and its API contract. Test the OpenClaw agent's integration with the prediction service and Gitea. Verify error handling, safety controls and human-approval requirements. Run reproducible evaluations comparing the base and fine-tuned models. Ensure the implementation, configuration files and technical documentation accurately reflect the working system. Expected Deliverables A working, independently tested AI/ML subsystem. Correctly configured model inference endpoints. Reproducible test results and evaluation evidence. Updated repository configuration and relevant technical documentation. A clear explanation of all changes made. A walkthrough of the system so I can independently operate, troubleshoot and demonstrate it.