Audio Engineer (Speech / TTS Audio Specialist) – French

$50/hrrate as listed

Clean, quality-check, and evaluate French-language voice recordings destined for TTS and speech AI training pipelines. Work is batch-based and spec-driven, combining audio restoration in tools like iZotope RX with critical listening judgments about French pronunciation and naturalness.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work involves

You receive batches of French speech recordings — voice talent sessions, scripted prompts, or scraped corpora — and bring them to a defined technical specification. Day to day that means silence trimming, breath handling, de-clicking, de-essing, and targeted noise reduction, followed by QA passes for clipping, distortion, residual artifacts, and loudness consistency across the batch. Alongside the engineering, you make linguistic calls: flagging mispronunciations, unnatural prosody, liaison errors, regional inconsistencies, or reads that would teach a model the wrong thing.

The judgment layer is what distinguishes this from general audio post. A take can be technically clean and still unusable for TTS because the delivery is stilted, the mic distance shifts mid-utterance, or the speaker's realization of a word doesn't match the transcript. You'll be asked to document why a file was rejected, not just that it was, and your reasoning feeds back into voice talent direction and spec revisions.

What the platform screens for

Verified French fluency at native or near-native level — screens probe for the ability to describe specific phonetic and prosodic phenomena, not just conversational ability.
Speech editing experience specifically, distinct from music production. Expect questions about dialogue and voiceover workflows, not mixing or mastering.
Tool fluency in iZotope RX, Pro Tools, Audacity, Reaper, or comparable restoration software, including which module you'd reach for given a described defect.
Consistency at volume — evidence you've worked structured, high-throughput pipelines where the same decision must be made identically across hundreds of files.

Logistics

Fully remote and largely asynchronous, with batches assigned and returned on defined turnarounds. Hours are typically flexible and part-time, ramping with project demand; some engagements request a minimum weekly commitment. Pay observed at $50/hr for this listing — Mercor sets rates per project and per candidate, so treat this as a reference point rather than a guarantee. You'll need your own DAW licenses, reliable monitoring (accurate headphones or nearfields), and a quiet listening environment.

What they look for

Native or near-native French with strong critical listening for phonetics and prosody
Hands-on speech and voiceover editing experience (not music-only)
Proficiency with iZotope RX, Pro Tools, or equivalent restoration software
Working knowledge of speech audio specs: loudness targets, clipping, noise floor, sample rate
Experience with high-volume, spec-driven batch workflows
Familiarity with TTS, speech ML, or dataset annotation pipelines (preferred)

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Are you a native French speaker, or have you lived in a French-speaking country for at least five consecutive years and use French daily in professional settings?
Native or near-native French fluency · required
Do you have at least two years of hands-on experience editing speech or voice recordings (not music production) using professional audio software?
Speech/voice recording editing experience · required
Do you have working experience with at least one professional audio restoration or editing tool such as iZotope RX, Pro Tools, Adobe Audition, or similar software?
Audio restoration tool proficiency · required
Have you worked on a project involving processing, QA-checking, or annotating batches of 100+ audio files or comparable high-volume structured audio work?
High-volume dataset or structured workflow experience · preferred

0 of 4 answered

Included with this listing

Interview prep — 11 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether candidate has hands-on experience with speech audio (not music), mastery of restoration tools, and ability to diagnose French-specific defects like liaison errors or unnatural pausing.
A strong answer: Must name specific tools (iZotope RX, Pro Tools), describe concrete speech issues encountered (breath noise, clipping, inconsistent loudness, or pronunciation artifacts), and explain the technical steps taken to resolve them—not vague cleanup but targeted intervention.
Where people lose it: Candidate conflates music production with speech editing, vaguely describes 'cleaning up audio' without naming tools or specific defects, or cannot articulate why the fix mattered for French speech quality.