Los Angeles Metropolitan Area
I am experienced in data annotation & labelling, QA annotation, data entry, and audio transcription. As a heritage speaker of Vietnamese, I have researched, designed, and tested the efficacy of my own generative LLM-based chatbots for heritage language learning for my culminating Master's project. My linguistic interests include conversation design, voice-user interface design, and sociolinguistics.
• Label, quality-check, and audit datasets to improve model accuracy for Meta’s commerce AI chatbot • Apply linguistic expertise and conversation design practices to evaluate conversational data in English, Vietnamese, and Tagalog • Achieved and maintained labeling accuracy, ensuring high-quality training data for AI models • Collaborate with engineers and researchers to refine labeling guidelines and boost model performance • Flag edge cases, patterns, and ambiguities to inform product and research teams • Support cross-functional teams in scaling labeling processes and optimizing AI/ML workflows • Deliver data-driven feedback that reduce annotation ambiguity and increase model training efficiency
• Assess prompt clarity, ambiguity, and potential harm • Determine if prompts are seeking information or time-sensitive • Utilize enhanced fact-checking to ensure AI-generated response accuracy • Evaluate responses for naturalness, comprehensiveness, and adherence to guidelines
• Employ linguistic expertise to score and improve upon existing prompts (questions) and AI-generated responses across a range of topics in English for client's LLM • Reference online resources to rephrase and write cohesive, accurate, responsive, and sometimes empathetic answers to prompts • Test (QA) and label other annotators' work based on guidelines • Correct incorrect prompt responses using natural language and examples • Evaluate AI model responses to prompts through scoring, ranking, A/B testing, etc.
• Provided instructional support for the Department of Advanced Studies in Education and Counseling • Graded assignments and provided feedback according to provided rubrics • Created annotated bibliography on emerging instructional materials for faculty
• Employed knowledge of acoustic and articulatory phonetics to accurately transcribe audio files • Collaborated with fellow transcribers to extrapolate linguistic patterns from data • Interpreted waveforms and acoustics via Praat software • Labelled and sorted transcribed audio via Microsoft Office Suite • Followed the International Phonetic Alphabet (IPA) for proper transcription of detailed utterances • Results in improving development of automatic speech recognition software (ASR) for children's speech pathology assessments