Henry H.

Key Skills

Software

Don't disclose

Top Subject Matter

AI-generated image evaluation and quality assurance

LLM alignment (RLHF/SFT)

response evaluation

Top Data Types

Image

Text

Video

Top Task Types

RLHF

Action Recognition

Segmentation

Freelancer Overview

Image AI Trainer, RemoExperts. Core strengths include Don't disclose, Internal, and Proprietary Tooling. Education includes Master of Philosophy, University of Ghana (2024) and Bachelor of Science, University of Ghana (2023). AI-training focus includes data types such as Image, Text, and Video and labeling workflows including Evaluation, Rating, and RLHF.

Labeling Experience

Image Annotation Specialist (within Advanced AI Data Trainer role), Invisible Technologies

ImageSegmentation

Annotated 25,000+ sports images with human pose estimation and field alignment (homography) to support computer vision training. Labeled 10,000+ cattle farm images identifying equipment, humans, vehicles, rails, and animal poses to create structured datasets. Reduced annotation errors by 25% through iterative feedback loops and refinement of labeling guidelines with quality analysts. • Annotated 25,000+ sports images (pose estimation and field alignment via homography). • Labeled 10,000+ cattle farm images (equipment, humans, vehicles, rails, animal poses). • Achieved 98% accuracy while meeting deadlines and SOP requirements. • Reduced annotation errors by 25% through iterative QA feedback and guideline updates.

2024 - Present

Video Annotation Specialist (within Advanced AI Data Trainer role), Invisible Technologies

VideoAction Recognition

Labeled 500+ short videos for action recognition and related visual attributes including camera details and object detection tasks. Annotated 2,000+ cow welfare videos covering behavior, movement, safety, transport, and event classification. Maintained annotation quality with pixel-level accuracy expectations for computer vision models. • Labeled 500+ short videos for action recognition, camera details, and object detection. • Annotated 2,000+ cow welfare videos across multiple behavior/safety/transport categories. • Focused on accurate labeling suitable for computer vision training. • Operated under SOPs while adapting to evolving requirements and maintaining high accuracy.

2024 - Present

Advanced AI Data Trainer, Invisible Technologies (USA)

TextRLHF

Annotated and labeled 1,000+ multi-modal datasets (text and images) to support RLHF and supervised fine-tuning for human preference alignment. Evaluated model responses using rubrics on clarity, factual accuracy, and guideline adherence to enable data-driven model optimization. Developed and produced structured prompts to train AI systems for precise, context-aware STEM education outputs. • Supported RLHF and SFT by labeling 1,000+ datasets (text, images). • Scored and annotated responses from different AI models using rubrics (clarity, factual accuracy, adherence). • Created 600+ structured prompts (multiple-choice Question-Step-Response framework) for STEM outputs. • Improved QA outcomes through iterative feedback loops and guideline refinement collaboration.

2024 - Present

Image AI Trainer, RemoExperts

Don't discloseImage

Trained and evaluated AI-generated images by rating 1,000+ outputs with 98% accuracy using detailed evaluation rubrics. Assessed each output across instruction following, image consistency, visual quality, and AI-detection ("AI-ness"). Supplied structured justifications to document scoring rationales and support model performance improvements. • Rated 1,000+ AI-generated image outputs with 98% accuracy. • Evaluated four dimensions: instruction following, image consistency, visual quality, and AI-ness. • Documented structured justifications for scoring decisions. • Supported large-scale image evaluation workflows for advanced image generation system development.

2025 - 2026

Computer Vision

ImageObject Detection

Annotated 25,000+ sports images (basketball, soccer, American football), including human pose estimation and field alignment (homography). Labeled 10,000+ cattle farm images, identifying equipment, humans, vehicles, rails, and animal poses. Annotated 2,000+ videos on cow welfare, including behavior, movement, safety, transport, and event classification.

2024 - 2026

Education

Work History