Analyst (Speech and Audio AI evaluation)

Innodata Inc.; Uttar Pradesh; On-site; INR 600,000 - 900,000; Full time

Job summary

Innodata Inc. is seeking Speech & Audio AI Evaluation Specialists to assess state-of-the-art S2S, TTS, and live voice agents in our India GCC. You will judge naturalness, prosody, and cadence across global English dialects, ensuring high-quality benchmarks.

Responsibilities include blinded pairwise ratings, multi-modal validation, and detailed documentation of phonetic artifacts. A C2 level fluency and formal auditory training are required for success.

Qualifications

• 2+ years of dedicated international voice process experience in a Captive / In-House GCC.
• C2 near-native English proficiency with deep conversational nuance mastery.
• Formally trained auditory ear for vocal inflections and articulation artifacts.

Responsibilities

• S2S & TTS naturalness scoring across intelligibility and cadence.
• Benchmark paralinguistic attributes like empathy, tone inflection, and urgency.
• Conduct blinded pairwise audio preferences with detailed rationales; verify multi-modal synchronization.
• Document timestamped acoustic anomalies and artifacts during evaluation.

Skills

C2 English proficiency

Auditory acuity

Voice process experience

Multilingual ability

Education

Linguistics / Phonetics background

Tools

Acoustic annotation tools

Job description

ROLE OVERVIEW & OBJECTIVE

We are looking for exceptional Speech & Audio AI Evaluation Specialists with genuine in-house Global Capability Centre (GCC) / Captive international voice experience to evaluate state-of-the-art Speech-to-Speech (S2S), Text-to-Speech (TTS), and real-time conversational voice agents. Unlike traditional transcription or BPO roles, this specialist position demands a highly trained auditory ear to evaluate prosody, cadence, phonetic accuracy, emotional steering, and paralinguistic nuance across global English dialects. Candidates must possess C2 level near-native fluency to execute rigorous human-preference and synchronization benchmarks.

KEY RESPONSIBILITIES & CORE WORKFLOWS

S2S & TTS Naturalness Scoring: Evaluate live Speech-to-Speech and neural Text-to-Speech outputs across intelligibility, rhythm, and conversational cadence.

• Paralinguistic & Emotion Steering Evaluation: Benchmark how effectively models express nuanced vocal attributes including empathy, hesitation, tone inflection, irony, and situational urgency.
• Pairwise Audio Preference Ratings: Conduct blinded, head-to-head auditory preference evaluations between candidate audio completions, providing detailed perceptual rationales. Multi-Modal data validations: Verify multi-modal synchronization, lip-sync alignment, and acoustic scene consistency for video dubbing and avatar-driven speech models.
• Accent & Dialect Calibration: Apply standardized rubric metrics uniformly across North American, British, Australian, and international English accents without regional bias.
• Defensible Auditory Documentation: Document timestamped acoustic anomalies, unnatural vocal artifacts, metallic distortion, and hallucinated phonetic segments.

CANDIDATE PROFILE & QUALIFICATIONS

Mandatory Requirements

• Experience: 2+ years of dedicated international voice process experience exclusively within a Captive / In-House Global Capability Centre (GCC) serving US/global client bases.
• Language Mastery: C2 near-native English proficiency with absolute mastery over conversational nuances, colloquialisms, idioms, and tonal dynamics.
• Acoustic Acuity: Formally trained auditory ear for subtle vocal inflections, pitch modulation, cadence shifts, and articulation artifacts.

Preferred Qualifications

• Linguistics & Phonetics: Academic coursework or practical background in phonetics, phonology, auditory acoustics, or speech-language pathology.
• Speech QA Background: Prior professional experience in voice quality analytics, acoustic data annotation, or speech synthesis evaluation.
• Acoustic Acuity: Formally trained auditory ear for subtle vocal inflections, pitch modulation, cadence shifts, and articulation artifacts.
• Multilingual Ability: Additional fluency in major European, Asian, or Latin American languages to support cross-lingual speech benchmarks.