David AI builds and licenses high-quality voice and audio datasets to the world's leading AI labs. We're a ~50-person Series B team running data collection across ~20 language markets, spanning conversational, expressive/emotional TTS, and transcription data.
The Role
You'll be David AI's in-house audio expert. When a contributor's recordings sound off, a program needs a recording spec, or we're setting up a studio session, you're the person who knows exactly what's wrong and how to fix it. You'll set the audio standards everyone else works from, help contributors and vendors get their setups right, and run the hardware and studio side of our in-person recording work. You'll report to the Head of Operations.
Be the go-to audio authority for Ops, Engineering, and ML, answering the "why does this sound wrong?" questions that nobody else can
Define audio quality standards and recording specs (noise floor, clipping, reverb, levels, mic placement, sample rate/format) for each program
Help contributors diagnose and fix their home recording setups, including mic choice, placement, gain staging, room treatment, and interface/software settings
Write clear setup guides, onboarding checks, and troubleshooting docs that non-experts can follow in any of our ~20 language markets
Lead a team and calibrate them to the company's audio standards. Listen to problem batches, identify root causes (environment, hardware, settings, processing), and recommend fixes
Define acoustic targets for processed, synthetic, or AI-generated audio, and judge whether output meets them
Work with ML researchers to identify and define audio events and acoustic quality issues (What counts as “muffled” audio? How do we define “robotic artifacts”?).
Turn these definitions into actionable guidelines for our expert audio QAs to label internal audio data.
Manage the QA team in a rapid iteration loop of labeling audio samples, getting feedback from ML researchers, and refining the annotation process to deliver high-quality training and eval sets.
Conduct quality audits on annotation and post-processing models before they land in production.
7+ years in audio engineering, recording, or production (studio, broadcast, podcast, post, or voice work)
A highly trained ear for subtle differences across voices, mics, rooms, and signal chains
Deep hands-on knowledge of mics, interfaces, preamps, acoustics, and room treatment
Strong DAW skills and fluency with loudness/metering, EQ, dynamics, noise reduction, and restoration tools (e.g., iZotope RX)
Ability to troubleshoot remotely and explain fixes clearly to non-technical people
Comfortable in a fast-moving startup: ambiguous problems, shifting priorities, high ownership
Nice to have
Experience with speech or voice data for AI (TTS, ASR, SFT projects) at a lab or data vendor
Experience recording voice actors or directing dialogue sessions
Experience supporting remote or distributed recording at scale
Scripting basics (Python, ffmpeg, SoX) for batch checks or processing
Multilingual
