Horizontal

AI-Driven Music Therapy System

Deployed in Henan University of Chinese Medicine

Location: China

Period: 2025 – present

As overall technical lead, architected a low-latency, closed-loop Music Therapy system which is deployed at Henan Medical University, Peking University International Hospital, and the Heilongjiang Provincial People’s Court. It streams high-fidelity EEG and audio, performs robust artifact rejection and feature extraction, and fuses neural and acoustic signals with a Transformer to drive a moment-to-moment generative music controller. Added 30 fps facial-affect estimation and voice-prosody analysis, then designed reliability-aware multimodal fusion so the controller adapts in real time and quantifies therapeutic response.

Real-Time Digital Human Dialogue System

Deployed in China Post Group,Heilongjiang Provincial People's Court

Location: China

Period: 2025 – present

Architected and executed a modular, ultra-low-latency pipeline for avatar rendering and full-duplex A/V via WebRTC, implementing an asyncio-based streaming stack that reliably sustains 30 fps video and 16 kHz audio for a tight perception-action loop. The system has been demonstrated and deployed at the China Post Group.