⚙ Tune
Infrastructure Engineer

Infrastructure
for agent-first
audio systems.

Infrastructure engineer building self-hosted, verifiable pipelines — a multi-site homelab, Cloudflare-tunneled services, and agent-orchestrated tooling. I automate the toil and keep humans the authority.

Contact Deck
Education NYU Steinhardt B.M. Music Theory and Composition for Film and Multimedia 2015–2019 · New York, NY
Specialization IRCAM Sound Spatialization and Synthesis 2017 · Paris, France
Certifications Wwise · UE5 CompTIA A+, CompTIA Network+, CompTIA Security+, Q-SYS, Dante Certification in progress 2 earned · 5 in progress
01

About

Infrastructure engineer building agent-first systems for audio production and beyond. I design self-hosted, verifiable pipelines — from a multi-site Tailscale homelab and Cloudflare-tunneled services to LUFS Workchain, a YAML-driven, agent-orchestrated audio-processing engine. I automate the toil and keep humans the authority, with a bias toward correctness that's proven, not assumed.

02

Experience

2024 – Present · Remote
LUFS Audio

Founder & Infrastructure Engineer

  • Architected LUFS Workchain — a YAML-driven, agent-first audio-processing engine with three interfaces (Bash engine, Node CLI, Python FastMCP server) sharing one parser so they never diverge
  • Built its verification/certification pipeline (content-hash catalog + ed25519 signatures); components promote unverified → verified → certified
  • Operate a multi-machine, two-site Tailscale homelab self-hosting LiteLLM, Supabase/PostgREST, n8n, Vaultwarden, and Gotenberg behind Cloudflare tunnels
  • Ship static sites on Cloudflare (Workers, R2) with GitHub Actions CI/CD
  • Designed a multi-agent ecosystem (cloud + local daemons) sharing a versioned knowledge base over the GitHub API
2022 – 2024 · San Francisco, CA
Hinge Health

Sound Designer (Contract) — Audio Systems & Tooling

  • Built custom audio implementation pipelines (Wwise + Unity) and modular, data-driven sound systems for apps serving 70k+ daily users
  • Stood up an A/B testing framework that drove an 18% increase in session duration
  • Authored scalable guidelines letting engineering teams self-serve without an audio-team bottleneck
03

Education

2015–2019 · New York, NY
NYU Steinhardt

B.M. Music Theory and Composition for Film and Multimedia

2017 · Paris, France
IRCAM

Sound Spatialization and Synthesis

04

Certifications

Certified Wwise Certification Audiokinetic
Certified Unreal Engine 5 School of Video Game Audio
In progress CompTIA A+ ACI Learning
In progress CompTIA Network+ ACI Learning
In progress CompTIA Security+ ACI Learning
In progress Q-SYS QSC
In progress Dante Certification Audinate
05

Selected projects

Public repo · Python

ffmpeg-mcp-server

An MCP server giving AI assistants safe, batch FFmpeg access — every command runs in a Docker container with networking disabled, against a whitelist validator, to a standardized 48 kHz / 32-bit output.

View on GitHub → Public repo · Python

markov-chain-midi-analysis

Chord- and note-level Markov models over a MIDI corpus. Rewritten after four defects in the original run were found and each reproduced under test; 31 contract tests now encode them.

View on GitHub → Public repo · C++ / Daisy Seed

echo-bridge

Partitioned-convolution reverb firmware for a guitar pedal — a 64-point FFT for early reflections, 1024 for the tail, impulse responses loaded from USB.

View on GitHub →
06

Skills

Infrastructure

Docker/Compose · Tailscale · Cloudflare (Tunnels, Workers, R2) · Linux self-hosting

Languages

Python · Bash · JavaScript/Node · Lua

Automation & Agents

n8n · MCP/FastMCP · LiteLLM · agent orchestration

Data & Services

Supabase/PostgREST · YAML pipelines · GitHub Actions CI/CD

Security & Correctness

ed25519 signing · content-hash verification · Vaultwarden secrets

07

Get in touch

Open to infrastructure and platform-engineering roles — email is the fastest way to reach me.

daniel@lufs.audio  ·  (714) 616-1558

Email GitHub LinkedIn Blog ⤓ Download deck (PDF)