Nathan Harmon

Founder · Engineer · Product

Nathan spent the last decade building signal-processing systems for defense and commercial teams, most recently RF geolocation and device identification at Northrop Grumman. He now leads Diffio end-to-end: training denoising models, building cloud GPU pipelines, and shipping simple tools creators can rely on.

Strengths
  • Speech restoration ML
  • Cloud GPU inference
  • Dataset & evaluation engineering
  • RF signal processing (background)
Focus
Making clean audio effortless for scrappy teams and solo creators
Base
Colorado Springs, United States
Email NathanOpen the contact formJoin the Discord

Currently shipping

  • Cloud GPU batch pipeline for speech cleanup and mastering (no real-time)
  • Workflow helpers for podcasters and small teams (upload → process → consistent outputs)
  • Model research for tough environments: mobile capture, reverberant rooms, low SNR

Looking for

  • Partners who know distribution, marketing, or community in creator audio
  • Feedback from podcasters and indie teams; share your toughest clip
  • Warm introductions to podcast networks, indie studios, and media teams that need easier post-production

Why it matters

Creating studio-quality sound is hard: good mics, quiet rooms, and audio know-how are expensive. Diffio closes that gap so you can record on a phone and publish audio that sounds intentionally produced. Lower barriers → more real conversations.