Kamanguya Wamae

Key Skills

Software

Remotasks

Clickworker

Mindrift

OneForma

Scale AI

Top Subject Matter

Speech Recognition

Natural Language Processing

Multilingual Linguistics

Top Data Types

Audio

Image

Text

Top Task Types

Transcription

Object Detection

RLHF

Data Collection

Text Summarization

Text Generation

Freelancer Overview

Detail-oriented Data Operations Specialist and AI Tutor with extensive experience in training Large Language Models (LLMs) and speech recognition systems. Proven track record at Outlier and Remotasks in providing high-fidelity annotations, transcriptions, and linguistic evaluations. Native English speaker with a deep understanding of regional accents and cultural nuances. Expert at using proprietary software to refine AI responses, ensuring natural prosody, rhythm, and accuracy in multilingual contexts. escalation backlogs. AI-training focus includes data types such as Audio and labeling workflows including Transcription.

Labeling Experience

AI Tutor / Data Contributor

RemotasksAudioTranscription

As an AI Tutor and Data Contributor at Outlier AI & Remotasks, I executed annotation tasks focused on speech recognition and natural language processing for high-priority AI projects. My responsibilities included transcribing diverse audio samples, ensuring linguistic nuances such as emotion and tone were accurately captured. I also provided feedback on AI-generated voice outputs and addressed potential bias related to regional and cultural variations. • Performed over 500 hours of audio transcription and annotation. • Applied RLHF practices to improve model speech-to-text accuracy by 90%. • Evaluated and provided qualitative analysis of AI-generated speech for naturalness and prosody. • Utilized proprietary and Remotasks platforms for data labeling in multilingual contexts.

As an AI Tutor and Data Contributor at Outlier AI & Remotasks, I executed annotation tasks focused on speech recognition and natural language processing for high-priority AI projects. My responsibilities included transcribing diverse audio samples, ensuring linguistic nuances such as emotion and tone were accurately captured. I also provided feedback on AI-generated voice outputs and addressed potential bias related to regional and cultural variations. • Performed over 500 hours of audio transcription and annotation. • Applied RLHF practices to improve model speech-to-text accuracy by 90%. • Evaluated and provided qualitative analysis of AI-generated speech for naturalness and prosody. • Utilized proprietary and Remotasks platforms for data labeling in multilingual contexts.

2024 - Present

Education

Work History