Danijela Blažević

Key Skills

Software

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Label Types

Audio Recording

Classification

Evaluation Rating

Prompt Response Writing SFT

Question Answering

Red Teaming

Text Generation

Freelancer Overview

I have professional experience as a Generative AI Annotator, providing structured training and evaluation data to improve model performance, alignment, and safety compliance. My work includes annotating and assessing prompts and model outputs, identifying failure patterns, policy misalignment, and edge-case behaviors across complex and sensitive domains. I have contributed to refining annotation guidelines, performing quality assurance on labeled data, and collaborating with cross-functional teams to ensure consistency, clarity, and high annotation standards. I also have experience evaluating model behavior in high-risk safety contexts, where nuanced judgment, contextual interpretation, and strict adherence to policy are essential. DANIJELA VURBIC My background in Psychology and Cyberpsychology, combined with Trust & Safety platform experience, allows me to bring a strong human-behavior lens to AI training data. I am particularly skilled at identifying subtle contextual signals, ambiguity, and policy gray areas, and at anticipating how real-world users — including vulnerable populations — may interact with AI systems. This enables me to produce high-quality annotations that support both model capability and robust safety outcomes.

Labeling Experience

Generative AI Annotator

Label StudioTextClassificationQuestion Answering

In my current role as a Generative AI Annotator, I produce high-precision training and evaluation data to support model alignment, performance, and safety compliance across multimodal systems. My work includes structured annotation of text, image, and audio inputs; detection of policy violations; and analysis of failure patterns in sensitive and high-risk domains. I have contributed to red-teaming style annotation in areas related to child safety and exploitation, helping identify model vulnerabilities and improve safeguard effectiveness. I also support guideline refinement, annotator calibration, and quality assurance processes, collaborating cross-functionally to ensure data integrity and that model behavior aligns with safety, risk, and policy standard.

In my current role as a Generative AI Annotator, I produce high-precision training and evaluation data to support model alignment, performance, and safety compliance across multimodal systems. My work includes structured annotation of text, image, and audio inputs; detection of policy violations; and analysis of failure patterns in sensitive and high-risk domains. I have contributed to red-teaming style annotation in areas related to child safety and exploitation, helping identify model vulnerabilities and improve safeguard effectiveness. I also support guideline refinement, annotator calibration, and quality assurance processes, collaborating cross-functionally to ensure data integrity and that model behavior aligns with safety, risk, and policy standard.

2023

Education

Work History