Finnish Voice AI Quality & Evaluation Specialist

TELUS Digital AI Data Solutions; Turku; On-site; EUR 42,000 - 64,000; Full time

Job summary

TELUS Digital AI Data Solutions is seeking a Finnish Speech & Audio AI Specialist to evaluate and validate speech AI systems, ensuring natural-sounding voices and accurate speech recognition. You will contribute to dataset validation, audio annotation, and quality assurance across TTS and ASR projects.

The role requires native Finnish (C1/C2+) and professional English (B1+), with hands-on experience using major DAWs and audio tools to ensure high-quality datasets for training and evaluation.

Qualifications

• Native-level Finnish required for pronunciation accuracy.
• Professional English required for technical instructions and documen- tation.
• Hands-on experience with professional audio tools and QA.

Responsibilities

• Evaluate AI-generated speech for naturalness, pronunciation, rhythm and emotion.
• Identify transcription errors and linguistic nuances in ASR outputs.
• Analyze audio with waveform and spectrogram tools to spot quality issues.
• Provide structured feedback to improve voice AI models.
• Prepare speech content for training pipelines and validation tasks.
• Support TTS/ASR/voice similarity projects and QA workflows.

Skills

Finnish (native)

English (professional)

Audio quality assessment

Education

Bachelor's degree in Audio Engineering/related field

2+ years in audio production or AI speech projects

Tools

Adobe Audition

iZotope RX

Pro Tools

Reaper

Audacity

Logic Pro

Cubase

Job description

TELUS Digital AI Data Solutions is seeking a Finnish Speech & Audio AI Specialist to evaluate and validate speech AI systems, ensuring natural-sounding voices and accurate speech recognition. You will contribute to dataset validation, audio annotation, and quality assurance across TTS and ASR projects.

The role requires native Finnish (C1/C2+) and professional English (B1+), with hands-on experience using major DAWs and audio tools to ensure high-quality datasets for training and evaluation.