Dmitriy Bulat

Key Skills

Software

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Classification

RLHF

Evaluation Rating

Freelancer Overview

I am a detail-oriented professional with a strong background in structured digital work and technical systems. While transitioning into AI data annotation and training projects, I bring proven experience in following detailed guidelines, maintaining consistency, and executing repetitive tasks with high accuracy. My background in web design and technical electrical work has developed my precision, focus, and ability to interpret written instructions carefully — all critical skills for data labeling and AI evaluation tasks. I am particularly strong in text classification, sentiment analysis, AI response evaluation, and content categorization. I work independently in remote environments, maintain sustained productivity (6+ hours daily), and prioritize quality control over speed. My disciplined workflow, reliability, and attention to detail allow me to contribute effectively to AI training, data annotation, and content moderation projects.

Labeling Experience

I Text Classification & Response Evaluation Project

CVATTextClassificationRLHF

Performed structured text annotation and AI response evaluation tasks simulating real-world LLM training workflows. The project involved text classification across multiple domains (finance, technology, healthcare, politics), sentiment analysis, and quality scoring of AI-generated responses. Applied multi-criteria evaluation methodology including accuracy, relevance, safety, and helpfulness scoring. Maintained strict adherence to annotation guidelines and consistency across edge cases. Ensured high labeling precision and internal quality control through systematic review of ambiguous samples. Project scope included 500+ annotated text samples across classification and evaluation tasks.

Performed structured text annotation and AI response evaluation tasks simulating real-world LLM training workflows. The project involved text classification across multiple domains (finance, technology, healthcare, politics), sentiment analysis, and quality scoring of AI-generated responses. Applied multi-criteria evaluation methodology including accuracy, relevance, safety, and helpfulness scoring. Maintained strict adherence to annotation guidelines and consistency across edge cases. Ensured high labeling precision and internal quality control through systematic review of ambiguous samples. Project scope included 500+ annotated text samples across classification and evaluation tasks.

2025

Education

Work History