Freelance AI Data Annotation (Alignment)

Perit AI; United States; Remote; USD 23,000 - 43,000; Full time

Job summary

Perit AI is hiring for a transcription alignment role. You will listen to audio and align each spoken word to its transcript position, correcting boundaries and transcription errors. The work is detail-heavy, quiet, and done remotely with flexible scheduling.

Strong English writing and comprehension are essential. No prior annotation experience is required; training and guidelines are provided. This is a remote, delivery-based role with payments per audio hour.

Qualifications

• Must be able to read and follow detailed guidelines.
• Attention to detail in transcription alignment.
• Willingness to work remotely with flexible scheduling.

Responsibilities

• Listening to audio and aligning each word in the transcript to where it is actually spoken.
• Correcting timing boundaries, missing or extra words, and transcription errors.
• Applying detailed annotation guidelines consistently across large volumes of work.
• Flagging unclear audio, edge cases, and anything the guidelines don't cover.

Skills

Strong English writing

Attention to detail

Good ear for language

Reading guidelines

Tools

Headphones

Job description

You'll listen to audio and align spoken words to their exact position in the transcript — correcting boundaries, catching mis-transcriptions, and making sure every word lands where it belongs. The work is detail-heavy and quiet. If you like careful, focused tasks and have a good ear for language, it's a strong fit. Fully remote and fully flexible. No fixed schedule and no minimum weekly commitment. You pick up tasks when they're available and deliver against a deadline — mornings, nights, weekends, whatever works for you.

What you'll be doing

• Listening to audio and aligning each word in the transcript to where it's actually spoken.
• Correcting timing boundaries, missing or extra words, and transcription errors.
• Applying detailed annotation guidelines consistently across large volumes of work.
• Flagging unclear audio, edge cases, and anything the guidelines don't cover.

What you'll need

A laptop or desktop computer. This work can't be done on a phone or tablet. Headphones and a quiet environment. You'll be listening closely to short segments of speech, often repeatedly. Strong English writing and comprehension. You should be comfortable reading detailed guidelines and picking up on subtle differences in wording. No prior annotation experience is required. Training and guidelines are provided.

Pay

Up to $17 to $31 per hour of audio output. Paid against delivery, with no minimum weekly commitment.

Listed rate: $17–31/audio hour.

Perit is a Y Combinator-backed speech-data company. Work is remote.