I have hands-on experience in data labeling and AI training data across multiple task types and platforms, including text, audio, and image-based work. My work has included RLHF, red teaming, prompt-response writing and SFT-style tasks, transcription review and correction, quality evaluation and rating, data collection, and bounding-box image annotation. I have evaluated AI-generated outputs for naturalness, accuracy, factual consistency, instruction adherence, and overall quality, compared alternative responses, written rationales to justify decisions, and contributed to the creation of high-quality example outputs for model improvement. I have also reviewed AI-generated transcriptions for orthography and completeness, corrected them when needed, and fully re-transcribed content when the AI output was too inaccurate to be usable.
What sets me apart is the combination of strong linguistic sensitivity and structured analytical judgment. As a native German speaker with strong English skills, I can work effectively in both languages, including multilingual and code-switching contexts, and I pay close attention to tone, nuance, phrasing, register, and cultural fit. In addition to text-based work, I have experience with audio tasks for LLM-related projects and image annotation for computer vision workflows, which gives me broader exposure than candidates who have only worked in one modality. I am comfortable following detailed guidelines, identifying subtle quality differences, analyzing error patterns, and contributing reliable human judgment across a wide range of AI training workflows.
Freelance AI Transcription Reviewer / Language QA Specialist
AudioTranscription
-Reviewed AI-generated transcriptions for orthography, completeness, and overall accuracy
-Corrected transcription errors and fully re-transcribed content when AI outputs were too inaccurate or incomplete
-Analyzed transcription error patterns and assessed issue severity
-Applied strong linguistic judgment in German and English audio-related tasks
-Reviewed AI-generated transcriptions for orthography, completeness, and overall accuracy
-Corrected transcription errors and fully re-transcribed content when AI outputs were too inaccurate or incomplete
-Analyzed transcription error patterns and assessed issue severity
-Applied strong linguistic judgment in German and English audio-related tasks
2026 - Present
Image Annotator
ImageBounding Box
- Annotated image-based data using bounding boxes for AI training tasks
- Marked and labeled relevant objects in images according to project guidelines
- Maintained accuracy and consistency across annotation tasks
- Followed detailed instructions to ensure high-quality computer vision training data
- Contributed to the preparation of structured visual datasets for machine learning models
- Annotated image-based data using bounding boxes for AI training tasks
- Marked and labeled relevant objects in images according to project guidelines
- Maintained accuracy and consistency across annotation tasks
- Followed detailed instructions to ensure high-quality computer vision training data
- Contributed to the preparation of structured visual datasets for machine learning models
2025 - Present
Freelance Prompt-Response Writer / AI Content Specialist
TextPrompt Response Writing SFT
-Created high-quality prompt-response pairs to support AI training and supervised fine-tuning workflows
-Wrote structured, high-quality example answers designed to demonstrate ideal model behavior
-Refined responses for clarity, naturalness, relevance, and instruction adherence
-Helped define what strong AI outputs should look like across different prompt types and task formats
-Applied strong linguistic judgment in German and English, with close attention to tone, phrasing, nuance, and overall response quality
-Created high-quality prompt-response pairs to support AI training and supervised fine-tuning workflows
-Wrote structured, high-quality example answers designed to demonstrate ideal model behavior
-Refined responses for clarity, naturalness, relevance, and instruction adherence
-Helped define what strong AI outputs should look like across different prompt types and task formats
-Applied strong linguistic judgment in German and English, with close attention to tone, phrasing, nuance, and overall response quality
2025 - Present
Freelance AI Quality Evaluator / Rater
TextEvaluation Rating
-Evaluated AI-generated outputs for naturalness, accuracy, relevance, instruction adherence, and overall quality
-Compared alternative responses and assigned ratings based on detailed evaluation criteria
-Wrote clear rationales to justify ratings and ranking decisions
-Reviewed outputs across text, audio, image, and other AI-related task formats
-Evaluated AI-generated outputs for naturalness, accuracy, relevance, instruction adherence, and overall quality
-Compared alternative responses and assigned ratings based on detailed evaluation criteria
-Wrote clear rationales to justify ratings and ranking decisions
-Reviewed outputs across text, audio, image, and other AI-related task formats
2025 - Present
Freelance AI Red Teaming Specialist
TextRed Teaming
Tested AI systems with challenging and adversarial prompts to identify weaknesses, edge cases, and failure patterns
Evaluated model robustness, safety, and instruction-following under difficult or ambiguous conditions
Documented problematic outputs and analyzed recurring failure modes
Contributed to improving model reliability through structured stress testing
Tested AI systems with challenging and adversarial prompts to identify weaknesses, edge cases, and failure patterns
Evaluated model robustness, safety, and instruction-following under difficult or ambiguous conditions
Documented problematic outputs and analyzed recurring failure modes
Contributed to improving model reliability through structured stress testing
