Summary
Zhizheng Wu is an associate professor and founder with 11+ years driving research and productization in speech and audio AI, currently leading Mel Lab and academic work at CUHK‑Shenzhen. He built widely used community resources—authoring the Merlin speech synthesis toolkit and curating ASVspoof datasets—and has organized major challenges (Blizzard, Voice Conversion, ASVspoof) that shaped evaluation standards in the field. His industry experience spans labs at Meta, Apple (Siri TTS), JD.com and Microsoft Research, where he translated deep learning research into production TTS, voice conversion, and anti‑spoofing systems. As an IEEE/ACM editor and technical‑committee member he bridges academic rigor and real‑world deployment, and he advises startups (Sanas) while founding Amphion to commercialize speech AIGC innovations. An often‑overlooked strength is his track record of creating open, reusable research infrastructure that accelerated reproducibility and competition in speech community benchmarks.
11 years of coding experience
13 years of employment as a software developer
PhD, speech synthesis, voice transformation, speaker recognition, PhD, speech synthesis, voice transformation, speaker recognition at Nanyang Technological University Singapore
English, Chinese