Language Data Annotator

Job Responsibility: This is a bilingual Voice AI Training project where you’ll help develop cutting-edge language technology by providing natural recordings in two languages. We’re looking for really advanced language speakers who can demonstrate clear, authentic accents in both. Your recordings will directly support the training and evaluation of advanced AI models for one of our key clients. This is the Accent and Voice Test for Project Lexicon by micro1. Here are the test details: Instructions

• Recording: Record a short audio clip in two languages: the first ~20 seconds in the first language, and the next ~20 seconds in the second language. It can be a bit longer if needed, but not excessively long.
• Language & Locale: Make sure to specify the correct regional variant of the spoken language (e.g., English UK, English US, Spanish MX, Arabic Gulf, etc.). There will be a drop-down box in the screening questions once you sign up.
• Content: The audio content can be about your hobbies, your profession, the weather, or any other natural topic. Please refrain from using any songs, poems our any phrases that it'd be possible for you to read to do this in order to be as natural as possible.

Accent and Naturalness

• Only native level speakers in both languages will be considered for the role.
• Please speak as naturally as possible. If someone hears the recording, it should sound as if you were born and raised in that country .
• If you're a language teacher that speaks correctly but without accent and/or speak the language with other accent, please make an effort to sound as close to a n ative speaker of the language being reviewed.

Audio Quality Guidelines:

• Recording should be done in quiet rooms using good quality devices, e.g., a USB headset. A built-in laptop mic is acceptable if it’s clear. You can use your cellphone to record, but please do so in a recording app (like the Voice Memos app on iPhones). The output of this should be a standard MP3 (.mp3) or WAV (.wav) file, do not send the file in formats such as FLAC, AIFF or AAC/M4A.
• There should be no obvious background noise or acoustic distortion.
• Avoid excessive or unnatural "dead-air" in your recording (>1 second) at the start, middle, or end. If this occurs, please re-record.
• Maintain consistent mic distance to avoid volume fluctuations.

This is the file that will be shared with the client and will define whether you will move forward in the process , so please be mindful of this when recording - do your best and best of luck!