Temidayo Olowojebutu

Key Skills

Software

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Top Label Types

Segmentation

Evaluation Rating

Audio Recording

Transcription

Freelancer Overview

With over four years of hands-on experience as an AI Voice Data Contractor, I have built a deep specialization in creating high-quality speech datasets for machine learning applications. At SpeechGen AI, I recorded more than 200,000 utterances across 15 phonetically diverse script sets designed specifically for Automatic Speech Recognition (ASR) model training. My work included delivering parallel recordings in four distinct accents with matched prosody for dialect generalization datasets, as well as executing emotional speech datasets spanning eight target emotions at 500 utterances each — data critical for affective computing research. I maintain a re-record rate of under 2%, reflecting the consistency and precision that AI training pipelines demand. What sets me apart is a unique combination of voice artistry and technical fluency in the specific requirements of AI data collection. I am well-versed in phoneme-balanced scripting, paralinguistic cue delivery (laughter, hesitations, filler words), prosody variation, and annotation-ready recording practices including metadata tagging and familiarity with IPA phonetic transcription. My medical background from the University of Abuja adds an analytical precision to my work, enabling me to follow complex linguistic and annotation protocols with accuracy. Whether delivering emotionally nuanced utterances or accent-controlled datasets, I bring both the vocal range and the technical discipline that AI speech systems require to learn natural, human-like patterns.

Labeling Experience

Coffee by Appen

AppenAudioSegmentationEvaluation Rating

The Coffee project was a large-scale Appen initiative focused on annotating naturally occurring human conversations to improve AI voice assistants and NLP systems. Working within the ADAP platform, I performed verbatim transcription capturing all stutters, fillers, and disfluencies, applied structured tags for non-speech sounds, conducted multi-speaker identification and turn-taking annotation, marked overlapping speech, and labeled emotional tone based on vocal delivery. The project was high-volume, involving hundreds of audio segments processed through a comprehensive multi-step annotation workflow. Quality was upheld by listening to each full segment before annotating, using playback tools to resolve unclear speech, maintaining strict consistency in tagging and transcription conventions, and following Appen's versioned internal guidelines throughout — ensuring all output met the standards required for model-ready AI training data.

The Coffee project was a large-scale Appen initiative focused on annotating naturally occurring human conversations to improve AI voice assistants and NLP systems. Working within the ADAP platform, I performed verbatim transcription capturing all stutters, fillers, and disfluencies, applied structured tags for non-speech sounds, conducted multi-speaker identification and turn-taking annotation, marked overlapping speech, and labeled emotional tone based on vocal delivery. The project was high-volume, involving hundreds of audio segments processed through a comprehensive multi-step annotation workflow. Quality was upheld by listening to each full segment before annotating, using playback tools to resolve unclear speech, maintaining strict consistency in tagging and transcription conventions, and following Appen's versioned internal guidelines throughout — ensuring all output met the standards required for model-ready AI training data.

2025 - 2025

Education

Work History