Keziah C.

Key Skills

Software

Appen

CVAT

Data Annotation Tech

Labelbox

Label Studio

Mindrift

OneForma

Remotasks

SuperAnnotate

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Audio Recording

Bounding Box

Emotion Recognition

Fine-tuning

Prompt + Response Writing (SFT)

Freelancer Overview

I am a detail-oriented AI training contributor with 2+ years of hands-on experience in audio transcription, text annotation, and evaluation of large language model (LLM) outputs. My work has focused on helping improve speech recognition systems, AI chat assistants, and emotion-aware video models. I have contributed to English audio transcription projects with precise speaker labeling and adherence to clean/verbatim guidelines. I’ve also evaluated AI-generated responses for coherence, appropriateness, and relevance, helping refine models for better user interaction. Additionally, I’ve labeled video content for action and emotion recognition, strengthening computer vision datasets. My academic background in Economics with a CPA minor gives me an edge in understanding complex and structured content. I’m fluent in English (C1+), proficient in a variety of labeling platforms (Labelbox, Appen, Scale AI), and consistently deliver high-quality, deadline-driven work.

Labeling Experience

Emotion and Action Recognition in Short Video Clips

LabelboxVideoClassificationEmotion Recognition

Labeled emotions and human activities in short video clips for use in emotion-aware and action-aware AI systems. Applied multi-class emotion tags (e.g., happy, confused, neutral) and action labels (e.g., walking, sitting, talking) to frames and sequences. Ensured accuracy by referencing visual cues and annotation guidelines. This project supported the training of affective computing and surveillance recognition models.

2025 - 2025

LLM Text Evaluation and Prompt Ranking

AppenTextQuestion AnsweringText Generation

Evaluated AI-generated responses for quality, relevance, logic, and tone across diverse prompts. Labeled ideal, acceptable, and poor completions based on detailed rubric criteria. Participated in supervised fine-tuning tasks (SFT) by providing human-written prompts and model responses. Helped improve chatbot alignment and usefulness for downstream users. Contributed over 500 evaluations with consistent rating quality.

2023 - 2024

English Audio Transcription for Speech Recognition

Scale AIAudioEntity (NER) ClassificationEvaluation/Rating

Worked on transcription of English audio clips for training speech-to-text models. Tasks included clean and verbatim transcription, accurate speaker labeling, and proper formatting of punctuation and context cues. Maintained high accuracy through adherence to detailed guidelines and review feedback. The project involved hundreds of short and long audio files in various accents and recording environments. Quality control was enforced through peer review and random audits.

2022 - 2023

Education

Work History