All Jobs

Detecta Language Agency is currently recruiting native Mexican Spanish speakers for a large-scale speech data collection project. The goal of this project is to collect and standardize 500 hours of natural full-duplex conversations in Mexican Spanish to support artificial intelligence and speech recognition technologies.

Participants will record natural two-person conversations in a quiet indoor environment. Conversations should be spontaneous and natural rather than scripted or read.

Project Requirements
• Language: Native Mexican Spanish
• Location: Mexico
• Age Range: 18–60 years
• Participants Required: 750+ speakers
• Gender Ratio: Male/Female approximately 1:1 (±5%)
• Recording Type: Two-person natural dialogue
• Recording Environment: Quiet indoor environment without echo
• Conversation Duration: 5–30 minutes per segment
• Maximum Recording Duration per group: Up to 2 hours

Recording Specifications
Audio recordings must meet the following technical standards:

• Sound card recording: 48kHz, 32-bit, mono
• Mobile phone recording: 24kHz, 16-bit, mono

Each recording session will produce three audio files:

1. Speaker A audio track
2. Speaker B audio track
3. Stereo reference file (used for alignment)

The two speaker tracks must be perfectly aligned in time and represent a natural dialogue between two speakers.

Conversation Guidelines
• Conversations should be natural and spontaneous
• Speakers will select topics from a provided topic list
• Reading scripts is not allowed
• Overlapping speech and interruptions are acceptable in natural conversation
• Content should avoid political, violent, or sensitive topics

Eligibility
Both experienced professionals and fresh native Mexican Spanish speakers are eligible to apply for this project.

Responsibilities
• Participate in natural conversation recordings
• Follow project recording guidelines
• Ensure clear and high-quality audio recordings
• Provide voice authorization for dataset usage