Finn O.

Key Skills

Software

LabelImg

Top Subject Matter

LLM Dialogue / Generative Text Evaluation

General Content & Media Annotation

Technology / Engineering

Top Data Types

Image

Text

Document

Top Task Types

Bounding Box

RLHF

Question Answering

Text Generation

Text Summarization

Evaluation/Rating

Point/Key Point

Red Teaming

Data Collection

Freelancer Overview

3 years of precision manufacturing process and QC experience, proficient in SOP development, FMEA/8D analysis, and Python/SQL data processing. Systematically mastered RLHF preference ranking, dialogue quality evaluation, and industrial visual annotation workflows. Specialized in manufacturing vertical domain, qualified for high-value tasks including industrial defect detection and predictive maintenance data construction. Committed to delivering accurate, consistent, and timely annotation and evaluation results.

Labeling Experience

AI Training and RLHF Evaluation Specialist

TextRLHF

Performed RLHF (Reinforcement Learning from Human Feedback) preference ranking and dialogue quality evaluation on AI-generated text samples. Applied evaluation rubrics to assess multi-criteria models and took part in large-scale batch evaluations and scoring. Delivered actionable feedback to improve LLM outputs and model alignment. • Scored LLM outputs for preference using defined rubrics • Conducted dialogue quality audits focusing on factuality, logical coherence, and user intent • Participated in annotation calibration sessions and quality reviews • Used Python for data validation, cleaning, and aggregation.

2022 - Present

AI Trainer/Data Annotation & Evaluation Specialist (Manufacturing domain)

LabelImgImageBounding Box

Worked on industrial defect annotation and quality evaluation using real-world manufacturing imagery data. Leveraged LabelImg and CVAT for creating bounding box annotations on gear machining and cutting tool datasets. Developed and reviewed annotation guidelines and ensured high inter-annotator agreement through SOP-based calibration. • Executed image annotation for industrial quality control tasks • Calibrated and audited annotators for consistency across multiple projects • Authored annotation guidelines and managed updates • Conducted manual data cleaning and batch processing using Python

2022 - Present

Education

Work History