? Upcoming Multilingual Code-Switching Speech Data Collection Project

Translingoworld Private Limited is looking for experienced speech data collection agencies, independent contributors, and vendor partners to support an upcoming Hebrew speech data collection project for Automatic Speech Recognition (ASR) model training.
We are seeking partners who can coordinate native speakers, manage recording sessions, and deliver high-quality speech recordings that meet the project's technical and compliance requirements.

? Project Overview
• Target language: Hebrew (Israel)
• Language code: he-IL
• Initial evaluation: One 15-minute sample recording
• Potential full-scale requirement: 200 hours of speech data

?️ Sample Recording Requirements
Each sample must meet the following specifications:
✅ Speakers: Two adults aged 18 or above, both native speakers of Hebrew as used in Israel and proficient in English.
✅ Natural conversation: Unscripted, spontaneous dialogue between the two speakers.
✅ Code-switching: Approximately 85% Hebrew and 10–15% English,



with natural code-switching throughout the conversation. Switching within sentences must be the primary pattern. English must not be read from a list or inserted on cue.
✅ Recording environment: A quiet, enclosed space with minimal background noise and no third-party voices.
✅ Audio format: Stereo WAV, with each speaker isolated on a separate audio channel. Mono or mixed-down recordings will not be accepted.
✅ Sample rate: Between 16 kHz and 48 kHz.
✅ Audio quality: Explicit, intelligible, distortion-free speech throughout the recording.
✅ Authenticity and consent: Genuine live human conversation only. Scripted or AI-generated dialogue is not permitted. Appropriate signed consent covering commercial ASR model training must be obtained before recording.

? Who Can Apply?
We welcome:
• Speech and language data collection agencies.
• Independent contributors who can coordinate compliant recordings with two eligible speakers.
• Vendor

📌 Hiring Hebrew Speech Data Collection Agencies (India)
🏢 Translingoworld Private
📍 India