SadiGroup

🌍 SadiGroup is Recruiting Strategic Vendor Partners for a Large-Scale Multilingual AI Speech Data Program

As part of our global AI data operations expansion, SadiGroup is seeking experienced vendors, language communities, and recruitment teams to support a multilingual speech data collection and annotation project across the following languages:

🇹🇭 Thai
🇵🇭 Filipino (Tagalog)
🇮🇩 Indonesian
🇲🇾 Malay
🇵🇹 Portuguese (Portugal)

Project Scope

🔹 500 valid hours per language
🔹 1,000 native speakers per language
🔹 30 minutes of effective speech per participant
🔹 Male/Female ratio: 1:1 (±5%)
🔹 Age range: 18–70 years old

Collection Requirements

✔ Native-born speakers of the target language
✔ Natural two-person conversations based on provided topics
✔ Conversation duration: 5–15 minutes per segment
✔ Dual-channel recording (one speaker per track)
✔ WAV format, 48kHz, 24-bit
✔ Quiet indoor environment with no background noise or echo
✔ Separate recording setup for each participant to avoid crosstalk

Annotation Requirements

✔ Audio transcription
✔ Utterance-level segmentation
✔ Speaker identification
✔ Quality validation

Quality Standards

🎯 Audio Qualification Rate: ≥99%
🎯 Transcription Accuracy: ≥98%
🎯 Annotation Accuracy: ≥95%
🎯 Speaker Identification Accuracy: ≥98%

Who Should Apply?

✅ AI Data Collection Companies
✅ Localization & Language Service Providers
✅ Community Leaders
✅ Recruitment Agencies
✅ Native Speaker Networks
✅ Vendors with proven multilingual sourcing capabilities

Long-Term Partnership Opportunity

SadiGroup is continuously expanding its global vendor ecosystem. Vendors who successfully deliver quality and scale on this project will be prioritized for future speech collection, transcription, annotation, RLHF, and multilingual AI data programs.

📩 If your organization can support one or more languages, we invite you to connect with our Project Management Team.

SadiGroup
Building the Human Intelligence Behind AI.

🚨 Hiring: Korean Elderly Audio Collection & Transcription Project (South Korea)

SadiGroup is looking for vendors, agencies, and recruitment partners in South Korea to support an upcoming Korean Elderly Speech Data Collection & Annotation Project.

📋 Project Overview

We are collecting authentic speech data from elderly Korean speakers to help improve AI speech recognition systems and reduce recognition errors for senior voices.

🎯 Participant Requirements

* Native-born Koreans
* Age: 65 years and above
* Currently residing in:
* Gyeonggi Province
* South Jeolla Province
* Born in the same province where they currently reside
* Total participants required: 10
* Gender ratio:
* 5 Male
* 5 Female
* Age distribution:
* 65–69: 2 participants
* 70–79: 6 participants
* 80+: 2 participants

🎙️ Recording Requirements

* Approximately 1 hour recording per participant
* Casual self-introduction and daily-life conversations
* Topics may include:
* Health status
* Daily activities
* Personal experiences
* Natural speaking style accepted, including pauses and hesitations
* Quiet indoor recording environment required

📂 Deliverables

* WAV format
* 16 kHz / 16-bit / PCM
* Participant metadata:
* Age group
* Gender
* Birthplace
* Current residence
* Signed consent form and ID verification required
* Transcription with timestamps for elderly speaker segments only

🤝 Who We’re Looking For

* Recruitment agencies
* Data collection vendors
* AI data service providers
* Teams with access to elderly Korean participants

📩 Interested?

Please send:

* Company introduction
* Relevant experience
* Available recruitment capacity
* Expected timeline
* Contact details

Apply or contact us:
📧 [email protected]
🌐 SadiGroup Website⁠
🔗 SadiGroup LinkedIn Page⁠