Paul M.

Key Skills

Software

Appen

Other

Telus

Top Subject Matter

Speech-to-text training

audio/video dataset annotation

Multilingual prompt localization and evaluation

Top Data Types

Audio

Text

Image

Top Task Types

Transcription

Classification

Prompt + Response Writing (SFT)

Data Collection

Freelancer Overview

Audio and Video Transcriber | Appen CrowdGen. Brings 8+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Appen, Other, and Telus. Education includes Master of Science, Kisii University (2024) and Bachelor of Science, Kisii University (2024). AI-training focus includes data types such as Audio, Text, and Image and labeling workflows including Transcription, Translation, and Localization.

Labeling Experience

Audio and Video Transcriber | Appen CrowdGen

AppenAudioTranscription

Performed high-accuracy transcription and annotation for audio/video datasets to support speech-to-text model training. Ensured phonetic and contextual precision through the use of advanced linguistic tooling. Maintained dataset quality requirements throughout labeling iterations. • Transcribed spoken content from audio/video inputs • Applied linguistic rules for phonetic and contextual accuracy • Annotated dataset elements required by downstream speech-to-text pipelines • Verified labeling consistency against quality criteria

2026 - 2026

Swahili Prompt Writer and Reviewer | Outlier

OtherTextClassificationTranscription

Crafted and reviewed complex Swahili language prompts to improve multilingual capability and cultural nuance. Conducted quality assurance testing to confirm localized content aligns with global model standards. Focused on improving prompt quality to enhance model outputs for Swahili use cases. • Wrote culturally nuanced Swahili prompts • Reviewed prompts for linguistic correctness and relevance • Ran QA checks to meet localization standards • Iterated prompt improvements based on evaluation results

2025 - 2026

Image Collection and Annotation | RWS

OtherImageClassificationTranscription

Curated and annotated image datasets to improve computer vision model accuracy. Used automated tagging tools to streamline data classification and reduce labeling latency. Prioritized precision and diversity so the resulting dataset supports robust model training. • Selected images to maintain dataset diversity • Applied image labeling for computer vision use • Used automated tagging to accelerate classification • Performed quality checks on annotated outputs

2025 - 2025

AI Annotator | Handshake

OtherTextClassification

Labeled and annotated complex inputs to support machine learning pipeline data integrity and quality. Followed structured annotation protocols to maintain consistent labeling across large-scale datasets. Performed quality checks to reduce errors and improve dataset reliability. • Applied annotation protocols to complex data • Ensured label consistency across dataset batches • Conducted integrity checks for labeled outputs • Supported downstream ML training through clean annotations

2025 - 2025

Prompt Response Writer | Telus

TelusTextPrompt + Response Writing (SFT)

Generated and evaluated diverse prompt responses to support large language model (LLM) training. Performed iterative human-in-the-loop testing to improve helpfulness and truthfulness of model outputs. Ensured response quality through repeated evaluation and refinement cycles. • Produced varied prompt-response pairs for LLM training • Assessed response quality for alignment and correctness • Conducted human feedback iterations to refine outputs • Documented evaluation outcomes to guide improvements

2024 - 2025

Show more

Education

Work History