About the Role
HireArt is recruiting an Audio AI Specialist to support the development of high-quality speech datasets for next-generation voice AI systems.
This role focuses on audio quality, annotation, and dataset validation — ensuring speech data is accurate, consistent, and usable for AI model training.
What You’ll Do
- Develop and apply audio quality standards for speech datasets
- Review and evaluate audio files for:
- Noise, distortion, clipping, echo, and recording issues
- Identify and flag:
- Segmentation errors
- Transcript mismatches
- Speaker labeling inconsistencies
- Perform annotation tasks:
- Transcription
- Timestamp validation
- VAD (voice activity detection)
- Diarization (speaker separation)
- Record high-quality speech from scripts
- Maintain consistency across audio, transcripts, and metadata
- Document edge cases and improve quality guidelines
- Collaborate with ML and research teams
Requirements
- Experience with audio AI datasets or speech workflows
- Hands-on experience with:
- ASR (speech recognition)
- TTS (text-to-speech)
- Speech-to-speech systems
- Strong ability to detect subtle audio quality issues
- Experience with:
- Transcription
- Segmentation / VAD
- Diarization
- Ability to produce clean recordings in a controlled environment
- Strong written communication skills
- High attention to detail
Bonus Skills
- Tools: Audacity, Praat, or similar
- Basic scripting (Python, SQL, Bash)
- Background in linguistics, phonetics, or voice work
- Experience with synthetic + real audio evaluation
- Multilingual or accent/dialect familiarity
Compensation
Role Details
- Remote (U.S. only, excluding CA & IL)
- Contract-based (long-term potential)
Why Join
- Work on cutting-edge voice AI systems
- Gain hands-on experience in speech datasets and model training
- Collaborate with AI researchers and engineers
- Build specialized skills in a fast-growing AI domain
About the Company
HireArt partners with leading companies to recruit talent for high-impact roles in AI, engineering, and data operations.