Kerryl

Key Skills

Software

Clickworker

Mercor

Anno-Mage

Argilla

Axiom AI

Data Annotation Tech

Top Subject Matter

Artificial Intelligence & Machine Learning

Ecommerce

Sports

Top Data Types

Image

Text

Computer Code Programming

Top Task Types

Bounding Box

Segmentation

Classification

Point Key Point

Question Answering

Text Generation

Text Summarization

Fine Tuning

Object Detection

Cuboid

RLHF

Polyline

Red Teaming

Transcription

Entity Ner Classification

Evaluation Rating

Polygon

Computer Programming Coding

Data Collection

Function Calling

Prompt Response Writing SFT

Freelancer Overview

My experience in AI training data centers on generating and validating high-quality Supervised Fine-Tuning (SFT) trajectories for Computer-Use Agents (CUAs) using frameworks like OpenClaw. I specialize in complex trajectory mapping, decomposing multi-step natural language instructions into precise, reproducible actions across file systems, terminal environments, and multi-tab browser navigations. My core focus is on producing "Gold Standard" training data by aggressively identifying negative trajectories, anticipating edge cases (such as permission errors or dynamic UI shifts), and documenting optimal "happy paths" to ensure models learn the most efficient, error-free logic. What sets me apart from standard data labelers is my background as a full-stack developer with over five years of experience engineering applications in Python, TypeScript, and PHP. I bring deep technical precision to the annotation process, ensuring that every command-line action or script execution within a training environment is structurally flawless. My hands-on experience with API integrations, DOM structures, and browser automation allows me to evaluate, debug, and correct agentic workflows at a granular, code-level depth, guaranteeing that the final datasets are 100% accurate and strictly reproducible for LLM training.

Labeling Experience

Education

Work History