Summary
Overview
Work History
Education
Skills
Websites
Hobbies and Interests
Languages
Core Competencies
Timeline
Generic

Caroline Jessen

København Ø

Summary

I'm a passionate AI Training Specialist and Data Annotation professional with a dedication to teaching AI to comprehend human context. I have an analytical mind and exceptional intuition, which helps me to make sure that complicated AI models are accurate and culturally sensitive, empathetic, and safe to use. I am able to work efficiently in cross-functional teams, and appreciate the collaborative working culture of the Danish tech scene. I'm excited to support creative teams to shape and implement responsible machine learning solutions with quality human in the loop interventions.

Overview

7
7
years of professional experience

Work History

AI Data & Annotation Lead

NordicTech Data Solutions
Copenhagen
08.2021 - Current
  • Human-in-the-Loop Supervision: Curated, formulated and analysed more than 50,000 text and image datasets using Scale AI to train Customer facing Conversational Agents.
  • Prompt Engineering and Response Evaluation: Implemented prompt engineering and assessed model-generated responses (ranking them in terms of accuracy, tone, and adherence to safety boundaries) using tools like Anthropic's Claude interface.
  • Model Quality Assurance: Reduced biases and the 'hallucination' occurrence in enterprise level applications by collaborating closely with the engineering teams on the back-end, resulting in a 30% drop in end-user complaint tickets.
  • Agile Coaching: Annotation data labeling training sessions were held every week and grading rubrics were developed to ensure a 98%+ consistency in data labelling tasks.

Digital Content Annotator / Junior Analyst

Institute of Marketing and Media
Roskilde
01.2019 - 07.2021
  • Visual Data Classification: Recognized and labeled visual data inputs with Labelbox for autonomous research and tech apps with highly complex bounding boxes and polygons.
  • NLP Processing: Linguistic data (Danish/English text structures) were cleaned and the semantic tagging was corrected.
  • Pipeline Efficiency: Produced standard operating procedure (SOP) documentation that helped standardize the process for passing annotations to the deep-learning team, saving the developers 15 hours per month in manual data cleaning.

Education

MSc - Information Technology (Cand.)IT in Digital Design & Interactive Technologies

IT-Universitetet i København (ITU)
Copenhagen, Denmark
01.2018

BSc - Communication and Digital Media

Aalborg University (AAU) – Copenhagen Campus
Copenhagen, Denmark
01.2016

Skills

  • AI & Data Annotation Tools
  • Scale AI
  • Labelbox
  • OpenAI API
  • ChatGPT
  • GPT-4
  • Fine-tuning
  • Hugging Face
  • GitHub Copilot
  • Cross-functional Collaboration
  • Ethical AI Compliance
  • Quality Assurance
  • Agile
  • Scrum Workflow Methodology
  • Independent Problem Solving
  • Analytical Thinking

Hobbies and Interests

Outside of work, I enjoy the sights of the city all year round in Nordhavn, where I bathe in the winter, and taking part in local running clubs around the Lakes. A person who strongly believes in the importance of balancing work and community life, I dedicate my time, as well as some weekends to coaching high school students in how to combine a creative approach with a modern career in technology.

Languages

  • Danish, Native level
  • English, Complete competent use of language in both listening and speaking

Core Competencies

Scale AI / Rapid: Prompt rating, formatting data and handling very large QA annotation sets., Labelbox: Knowledge of semantic segmentation, bounding box, and structuring of metadata., OpenAI API & ChatGPT (GPT-4): Implemented prompt engineering, output debugging, and LLM behavior tuning., Fine-tuning and output comparison with the constitution based AI frameworks, Anthropic Claude., Hugging Face / Datasets: Upload, clean and validate human evaluation datasets for machine learning testing., GitHub Copilot: Helping with script writing to streamline and automate data sorting pipelines., Cross-functional Collaboration, Ethical AI Compliance & Quality Assurance, Agile/Scrum Workflow Methodology, Independent Problem Solving & Analytical Thinking

Timeline

AI Data & Annotation Lead

NordicTech Data Solutions
08.2021 - Current

Digital Content Annotator / Junior Analyst

Institute of Marketing and Media
01.2019 - 07.2021

MSc - Information Technology (Cand.)IT in Digital Design & Interactive Technologies

IT-Universitetet i København (ITU)

BSc - Communication and Digital Media

Aalborg University (AAU) – Copenhagen Campus
Caroline Jessen