Sylvia G.

Key Skills

Software

No software listed

Top Subject Matter

Content moderation and licensing compliance for image-based media.

Bilingual text evaluation

translation quality checks

Top Data Types

Image

Text

Document

Top Task Types

Data Collection

Question Answering

Text Summarization

Freelancer Overview

I have more than three years of experience in helping with the collection and evaluation of data for artificial intelligence models through my employment at RWS Group and other data-related jobs. In my role as Meta AI Quality Evaluator and Google Bard Data Rater, I evaluated the responses generated by the Large Language Models using English and Swahili in terms of accuracy, fluency, consistency of semantics, relevancy, and adherence to prompts. I had a lot of experience evaluating multimodal data, which included text, image, audio, and video data.

Labeling Experience

Klarecon — Content Quality Assurance

Image

Provided content quality assurance by tracking sensitive or inappropriate images against platform editorial policies. Verified image accuracy and relevance relative to accompanying text content, and ensured licensing compliance for Creative Commons and stock media. Embedded certified images using CMS workflows such as WordPress to support compliant publishing. • Reviewed images for policy adherence and potential risk content • Checked alignment between images and their textual context • Applied licensing and copyright requirements during submissions • Used CMS tools to embed and certify assets

2025 - 2025

RWSGroup — Meta AI Quality Evaluator

Text

Evaluated Large Language Model (LLM) outputs in Swahili for accuracy, fluency, and prompt-response coherence. Assessed semantic similarity between sentences to confirm that responses matched the intended meaning. Applied consistent quality criteria to support improved language model reliability. • Reviewed LLM answers for correctness and naturalness • Measured semantic similarity between source and target sentences • Verified coherence between prompts and model responses • Reported quality findings to guide model improvement

2024 - 2025

RWSGroup — Google Bard Data Rater

Text

Performed LLM data rating work in Swahili by refining and fine-tuning model performance through detailed rating tasks. Assessed translation accuracy and evaluated semantic coherence between English and Swahili text pairs. Ensured quality of bilingual language outputs for downstream AI use. • Conducted Swahili rating tasks to improve LLM behavior • Evaluated translation quality and semantic consistency • Scored language pairs to measure coherence and accuracy • Supported iterative improvement of LLM performance

2023 - 2024

Education

Work History