William Daniel Schroll

Key Skills

Software

Appen

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Top Label Types

Question Answering

Translation Localization

Freelancer Overview

I am a detail-oriented Language Data Specialist with over 4 years of hands-on experience in audio transcription, data annotation, and AI training data preparation, with a strong focus on American English. Throughout my career, I have supported the development of NLP and ASR systems by delivering high-accuracy speech datasets for companies like Appen, Lionbridge AI, and iMerit. I have transcribed and annotated thousands of hours of audio, consistently maintaining a 98–99% accuracy rate and meeting project deadlines in fast-paced, remote environments. My expertise spans audio QA, speech and voice data labeling, accent and dialect identification, as well as linguistic data analysis. I am proficient in using a variety of data annotation platforms, including VGG Image Annotator (VIA), LabelImg, and Amazon SageMaker, and I have a strong background in translation, localization, and structured QA reporting. With my deep understanding of model training and data preprocessing, I have collaborated with data scientists and AI engineers to optimize datasets, ensuring consistency and high-quality outputs. My passion for linguistics and supporting AI and machine learning initiatives drives my commitment to delivering precise, reliable training data that enhances model performance and contributes to the success of advanced machine learning algorithms.

Labeling Experience

Audio Transcription & Speech Data Labeling for Healthcare ASR Systems

AppenAudioQuestion AnsweringTranslation Localization

I worked on a project for a healthcare client focused on improving the accuracy of an ASR system designed to transcribe doctor-patient interactions. My tasks included transcribing medical audio recordings, identifying speaker roles (doctor vs. patient), and ensuring that medical terminology, accents, and complex dialogues were accurately captured. This project involved labeling several thousand hours of audio data, with strict adherence to quality control measures, ensuring a 98-99% accuracy rate for transcriptions. Additionally, I performed regular audio quality assurance checks to flag any audio distortions or background noise that might impact the transcription process. My work helped enhance the speech-to-text model's ability to process and transcribe medical conversations more effectively.

I worked on a project for a healthcare client focused on improving the accuracy of an ASR system designed to transcribe doctor-patient interactions. My tasks included transcribing medical audio recordings, identifying speaker roles (doctor vs. patient), and ensuring that medical terminology, accents, and complex dialogues were accurately captured. This project involved labeling several thousand hours of audio data, with strict adherence to quality control measures, ensuring a 98-99% accuracy rate for transcriptions. Additionally, I performed regular audio quality assurance checks to flag any audio distortions or background noise that might impact the transcription process. My work helped enhance the speech-to-text model's ability to process and transcribe medical conversations more effectively.

2021 - 2021

Education

Work History