Audio,
and the systems
that run it.
I work where audio meets infrastructure — designing sound and building the systems that deliver it reliably at scale.
About
I work where audio meets infrastructure — designing sound and building the systems that deliver it reliably at scale. My background spans interactive audio and UX sound design for products people rely on every day, and self-hosted, agent-first pipelines that automate production without giving up correctness or control. Whether the question is how something should sound or how the signal, tooling, and delivery hold together, I've usually worked both ends of it.
Experience
Founder — Sound & Systems
- Sound design and audio implementation for tech clients including Hinge Health, xAI, Meta, and Roblox
- Built LUFS Workchain — a YAML-driven, agent-first audio-processing engine — and the verification pipeline behind it
- Self-host the production infrastructure: homelab, Cloudflare-tunneled services, CI/CD, and streaming (catalog.lufs.audio on Cloudflare R2)
Sound Designer (Contract)
- Designed the interaction-sound system for mobile health apps serving 70k+ daily users
- Built the Wwise/Unity implementation pipelines and an A/B testing framework (+18% session duration)
- Wrote guidelines that let engineering teams implement audio independently
Education
B.M. Music Theory and Composition for Film and Multimedia
Sound Spatialization and Synthesis
Certifications
Selected projects
lufs-recorder-pwa
Zero-install browser capture for audio and MIDI, live at rec.lufs.audio — AudioWorklet into hand-written 24-bit PCM, Web MIDI, waveform and piano-roll playback, and a take.json sharing the native recorder's schema. Gating checks prove a take; what the browser cannot know (xruns, channel identity) is reported as inferred rather than faked.
View on GitHub → Public repo · Pythonalbumdujour
Static-site generator that turns a listening log in Google Sheets into a published site, with a credential-safe build and a subtree-split deploy.
View on GitHub → Public repo · WebGLlufs-vectorscope-fog-demo
An audio-reactive vectorscope — a true Lissajous of left × right through a mid/side transform — veiled behind a shader fog. Self-contained, no build step.
View on GitHub →Skills
Interactive/UX sound design · Wwise, FMOD · Adaptive sound systems · Sonic identity
Cloudflare (Workers, R2) · Docker/Compose · Tailscale · Linux self-hosting
Python · Bash · JavaScript/Node · YAML pipelines · GitHub Actions CI/CD
Agent orchestration · MCP/FastMCP · n8n · Verifiable correctness
Get in touch
Open to roles across audio, infrastructure, and where they meet — email is the fastest way to reach me.
daniel@lufs.audio · (714) 616-1558