“Sound doesn't just reach us. It rearranges us.”

For a human, the first interface. For a machine, the next.

Contact

Awards, partners & institutions

More partners →

Me

Audio innovation, continuously delivered.

Content, software, hardware — twenty years in the field.

  • ExpressionTalktalk: a teleprompter tool even professional news anchors love, for sharper on-camera delivery (2026–); ‘Highlights Music Station,’ a showcase of China's independent musicians (2020–2022); host at China Radio International HIT FM (2006–2012); host & music coordinator at Phoenix U Radio, Hong Kong (2013–2014).
  • RelaxPILO® and CloudDream® sleep-aid audio (2015–).
  • PerformanceRun Baby Run® extended into an AI workflow — a cadence-music series where every step lands on the beat (2013–).

About → Writing →

On the record

All recognition →
Red Dot Award 2016 winner letter for PILO®
Red Dot AwardGermany · 2016
A' Design Award Silver certificate for PILO®
A' Design AwardItaly · Silver, 2017
Golden Pin Design Award 2018 certificate for PILO®
Golden PinTaiwan · 2018
Plus X Award certificate for PILO®
Plus X AwardGermany · Innovation
K-Design Award 2019 certificate for PILO®
K-Design AwardKorea · 2019
WIRED Germany magazine feature on PILO®
WIRED GermanyPILO® feature
Men's Health magazine feature on Run Baby Run
Men's HealthRun Baby Run® feature
Nike Shanghai Marathon Run Baby Run feature
Nike Shanghai MarathonRun Baby Run® soundtrack
PILO® — music-driven sleep pillow with companion app
BrookstonePILO® retail · China

Track record

Case studies in audio product design.

PILO® — music-driven sleep pillow with companion app

PILO®

The award-winning music sleep pillow.

Music-driven sleep system. Six international design awards. Featured in Wired. Sold through Brookstone (US), Amazon US & Japan, and Tsutaya Books, Japan.

Awards
Red Dot A' Design Golden Pin Plus X K-Design
Channels
Brookstone China JD Amazon US Amazon Japan
Red Dot entry A' Design page GeekPark feature
Talktalk product website snapshot

Talktalk

Elevate your expression, unlock your influence.

A Hong Kong TV news anchor: “It feels great — smooth and intuitive, never working against you. It keeps your attention entirely on what you're saying.”

System
Talktalk Real-time prompting Adaptive audio
Category axes
Expression Attention Performance
Open Talktalk Read recent writing

AURORA · Open source

Building audio AI from scratch.

A from-scratch course on audio AI — from a single sine wave to Whisper. 99 lessons, every algorithm hand-written in NumPy, no black-box libraries. Open core: Lesson 1 is free and open source; the full 99-lesson edition follows. I designed and directed the curriculum; built with AI-assisted pair programming.

Built from scratch
FFT MFCC Attention CTC RAG
Open core
Lesson 1 free 99 lessons NumPy only
Lesson 1 — free View on GitHub