Elosy G.

Key Skills

Software

Sama

Other

CloudFactory

Clickworker

CVAT

Data Annotation Tech

iMerit

Labelbox

Remotasks

Scale AI

Snorkel AI

V7 Labs

Toloka

Telus

Micro1

Top Subject Matter

Multilingual text classification dataset annotation (intent, sentiment, toxicity, safety).

Quality assurance for annotated datasets and AI outputs (linguistic and factual checks).

LLM content review and refinement (quality, coherence, cultural appropriateness).

Top Data Types

Text

Image

Video

Top Task Types

Classification

Text Generation

RLHF

Bounding Box

Polygon

Segmentation

Entity (NER) Classification

Point/Key Point

Cuboid

Polyline

Object Detection

Text Summarization

Question Answering

Red Teaming

Fine-tuning

Transcription

Evaluation/Rating

Computer Programming/Coding

Prompt + Response Writing (SFT)

Data Collection

Function Calling

Freelancer Overview

AI Data Annotator — Sama (SamaSource), Kenya. Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Sama and Other. Education includes Master of Science, University of Nairobi (UoN) (2025) and Bachelor of Science, Maseno University (2019). AI-training focus includes data types such as Text and labeling workflows including Classification, Evaluation, and Rating.

Labeling Experience

AI Trainer — Atlas (Remote)

OtherTextRLHF

Trained and evaluated LLM outputs for accuracy, clarity, and policy alignment. Delivered structured RLHF-style feedback to support model improvement. Ran prompt experiments to reduce failure modes and improve instruction-following behavior while maintaining annotation quality through edge-case documentation and checklists. • Assessed LLM responses for quality, correctness, and policy compliance. • Provided structured RLHF-style feedback to guide model updates. • Conducted prompt testing to improve instruction-following and reduce errors. • Used edge-case documentation and checklists to ensure consistent labeling quality.

2025 - Present

LLM Contributor — Outlier (Remote)

OtherTextText Generation

Reviewed and refined AI-generated content to improve coherence, correctness, and stylistic consistency. Standardized terminology and produced reference notes to support cultural appropriateness and clearer communication. Authored concise summaries and actionable recommendations intended for client use. • Evaluated AI outputs for coherence and correctness against intended meaning. • Improved stylistic consistency and standardized terminology for clarity. • Created reference notes/glossaries to enhance cultural relevance. • Produced client-ready summaries and recommendations based on reviewed content.

2024 - Present

Quality Reviewer — Innodata (Remote)

OtherText

Performed linguistic and factual QA on previously annotated datasets and AI outputs. Enforced style and taxonomy standards while documenting edge cases and rationales for alignment with client requirements. Collaborated with PMs and SMEs to ensure deliverables met agreed specifications. • Conducted linguistic QA including style/taxonomy compliance checks. • Verified factual correctness and consistency of AI outputs. • Documented edge cases and provided rationales to support decisioning. • Aligned outputs with client requirements in collaboration with stakeholders.

2022 - 2023

AI Data Annotator — Sama (SamaSource), Kenya

SamaTextClassification

Labeled multilingual and English datasets for intent, sentiment, toxicity, and safety using detailed annotation guidelines. Ensured outputs met taxonomy and guideline requirements while contributing to consistent labeling across examples. Participated in quality improvement by escalating ambiguous cases and supporting guideline updates. • Applied taxonomy interpretation and guideline enforcement to text inputs. • Worked to increase inter-annotator agreement through examples and clarifications. • Updated guidelines and quick-reference checklists to improve throughput and consistency. • Produced reliable, audit-ready annotations aligned to project standards.

2021 - 2023

Education

Work History