Senior Research Scientist at Shanda AI Research Tokyo.

I work on generative audio, voice LLMs, and multimodal foundation models.

With 7+ years of experience from research to production, I build product-ready speech and audio AI systems across expressive TTS, full-duplex speech systems, codec and tokenization design, large-scale multi-GPU training, and deployment.

• Website: zengchang233.github.io
• Google Scholar: Profile
• GitHub: @zengchang233
• LinkedIn: chang-zeng-5451a4191
• X (Twitter): @zzzzzc12

• Generative audio and voice LLMs
• Multimodal foundation models
• Speech and singing voice generation
• Speaker recognition and anti-spoofing
• Audio separation and enhancement

• 2025.09 - Present: Senior Research Scientist, Shanda AI Research Tokyo
• 2024.04 - 2025.08: Multimodal Generative AI Researcher, Li Auto
• 2023.09 - 2024.03: Speech ML Researcher (Intern), RevComm Inc.

• 2026.05: A Semantically Consistent Dataset for Data-Efficient Query-Based Universal Sound Separation accepted by ICML 2026.
• 2026.03: Released Speech Codec Probing from Semantic and Phonetic Perspectives on arXiv.
• 2026.01: DrivingScene and PAGS accepted by ICASSP 2026.
• 2025.12: Released Towards Interactive Intelligence for Digital Humans on arXiv.
• 2025.10: Critical Information Only accepted by IEEE TDSC.
• 2025.01: SonicSim accepted by ICLR 2025.
• 2025.01: A Benchmark for Multi-Speaker Anonymization accepted by IEEE TIFS.

• Speech Codec Probing from Semantic and Phonetic Perspectives (arXiv, 2026)
• A Semantically Consistent Dataset for Data-Efficient Query-Based Universal Sound Separation (ICML 2026)
• PAGS: Priority-Adaptive Gaussian Splatting for Dynamic Driving Scenes (ICASSP 2026)
• DrivingScene: A Multi-Task Online Feed-Forward 3D Gaussian Splatting Method for Dynamic Driving Scenes (ICASSP 2026)
• Critical Information Only: A Content Privacy-Preserving Framework for Detecting Audio Deepfakes (IEEE TDSC, 2025)
• SonicSim: A Customizable Simulation Platform for Speech Processing in Moving Sound Source Scenarios (ICLR 2025)
• A Benchmark for Multi-Speaker Anonymization (IEEE TIFS, 2025)