Caxton M.

Key Skills

Software

Other

Appen

Data Annotation Tech

Micro1

Mercor

LabelImg

Scale AI

Remotasks

Toloka

Snorkel AI

Top Subject Matter

LLM response evaluation

RLHF preference ranking

supervised fine-tuning data creation

Top Data Types

Text

Document

Image

Top Task Types

Classification

Transcription

Data Collection

Fine-tuning

RLHF

Computer Programming/Coding

Evaluation/Rating

Question Answering

Text Summarization

Text Generation

Object Detection

Polyline

Point/Key Point

Cuboid

Red Teaming

Polygon

Bounding Box

Segmentation

Entity (NER) Classification

Function Calling

Prompt + Response Writing (SFT)

Freelancer Overview

AI Trainer & Text Annotator — Mindrift (Toloka AI). Brings 8+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Mindrift (Toloka AI) and Other. Education includes Bachelor of Arts, University of Nairobi (2019). AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and Classification.

Labeling Experience

AI Trainer & Text Annotator — Mindrift (Toloka AI)

Text

Provided structured response-quality ratings for LLM outputs and evaluated outputs across multiple quality dimensions. Applied guidelines to assess factual accuracy, logical coherence, instruction adherence, fluency, and safety. Maintained consistent labeling quality with QA review cycles and escalation of edge cases. • Rated AI-generated text for accuracy, consistency, coherence, and safety. • Performed instruction-following assessments against user prompts. • Ensured inter-annotator agreement by following project ontology and style guides. • Coordinated with QA specialists during iterative review cycles.

2024 - Present

Communications & Content Officer — CABI (Centre for Agriculture and Bioscience International)

OtherTextClassification

Produced fact-checked, well-structured written content following strict editorial guidelines and content ontologies. Performed multi-stage feedback and refinement loops that align with iterative quality assurance workflows used in annotation projects. Supported practices of accuracy, consistency, and attention to detail required for gold-standard response authoring. • Wrote and revised content using structured editorial frameworks/ontologies. • Applied rigorous QA processes and participated in feedback loops. • Ensured factual correctness and consistency under guideline constraints. • Produced materials suitable for reference response authoring practices.

2022 - Present

Education

Work History