On-site AI Engineer: DGX, LLM Agents, Trading Automation

Private client in Central, Hong Kong seeks a hands-on AI engineer for on-site work in a home-office AI lab. You will commission an NVIDIA DGX Spark (Ubuntu ARM64), deploy local LLMs (vLLM/Ollama/SGLang with a web UI and RAG), build multi-model agent workflows and automations, and develop trading tools against the Interactive Brokers API — paper trading first; you build documented software only, and all live trading decisions remain with the client. Every visit includes structured 1-on-1 training plus plain-English runbooks so the client can operate everything independently. Hourly engagement: paid pilot task, then roughly 10-15 hours weekly with ongoing maintenance after. NDA, least-privilege access and security protocols apply. Fluent English required; IBKR API experience a strong plus.