Florina Curteanu

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Top Task Types

Object Detection

Question Answering

Text Generation

Text Summarization

Freelancer Overview

Over the 4 four years, I’ve combined my background as a researcher and analyst with hands-on work using various AI tools (since 2021), which naturally evolved into experience with data labeling and AI training data workflows. My PhD-level training in communication sciences and long-term work as a Senior Research Analyst mean I’m very comfortable designing coding schemes, building analysis grids, and applying consistent annotation criteria to complex, messy, real-world data. I’ve labeled and analyzed qualitative and quantitative data across domains (politics, media, HR, FMCG, pharma, IT, education, tourism, etc.), worked with questionnaires and large samples, and used SPSS and Python for validation, cleaning, and statistical checks. This combination has made me very detail-oriented, rigorous about guidelines, and attentive to bias, representativeness, and data quality, key aspects of reliable AI training data. What sets me apart is my experience with multimodal and social media data (text, visual, and video), my ability to design the methodology behind a labeling project (not just execute it), and my business-facing background. I’ve done advanced verbal, visual, and multimodal analyses of political debates, social media content, and audiences, turning coded data into insights and clear recommendations for clients and stakeholders. Since 2021, I’ve been integrating AI tools into these workflows, using them to accelerate research, test prompts and iterate on taxonomies

Labeling Experience

Research Coding

Scale AIVideoObject Detection

I also manually annotated thousands of segments from televised election debates, news clips, campaign visuals and social media posts in several European countries. I designed and refined coding schemes (topics, sentiment, stance, argument type, visual framing, speaker attributes), trained other coders and checked inter-coder reliability. Since 2021 I have integrated AI tools into this workflow to pre-classify data, test and adjust taxonomies, and run consistency checks, giving me hands-on experience with creating and validating high-quality training data for AI and LLM-style systems.

I also manually annotated thousands of segments from televised election debates, news clips, campaign visuals and social media posts in several European countries. I designed and refined coding schemes (topics, sentiment, stance, argument type, visual framing, speaker attributes), trained other coders and checked inter-coder reliability. Since 2021 I have integrated AI tools into this workflow to pre-classify data, test and adjust taxonomies, and run consistency checks, giving me hands-on experience with creating and validating high-quality training data for AI and LLM-style systems.

2010 - 2014

Education

Work History