Ann K.

Key Skills

Software

Scale AI

Other

Top Subject Matter

Healthcare & Life Sciences – Clinical, Regulatory & Medical Content

Legal – Contracts, Compliance & Regulatory Text

IT/Software & E-commerce – Product & Technical Content

Top Data Types

Image

Text

Audio

Top Task Types

Transcription

Evaluation/Rating

Data Collection

Red Teaming

RLHF

Fine-tuning

Question Answering

Object Detection

Text Generation

Classification

Freelancer Overview

Evaluation work for an LLM/model-development company (under NDA) across multimodal and text tasks. On the multimodal side, performed object-removal-and-restoration tasks — editing elements out of images and video, then prompting models to reconstruct them — and compared outputs across models to judge which performed better. On the text side, prompted and challenged models in Polish, rating response quality and naturalness as a native speaker, and ran multi-round consistency checks to assess whether a model held its answers stable across turns. Also conducted safety red-teaming, probing models to elicit PII and reporting pass/fail on whether the model resisted, to support guardrail improvement.

Labeling Experience

Multilingual AI Transcript Evaluation & QA (Medical + General)

AudioTranscription

Evaluate and quality-assure AI-generated transcripts across general and medical content, assessing accuracy, faithfulness to source audio, terminology correctness, and fluency of model output. I work with Polish (my native language), applying rigorous, detail-level linguistic review and consistent application of project guidelines. Beyond pass/fail scoring, I flag systematic model error patterns, recurring mistranscriptions, terminology drift, and domain-specific accuracy issues, particularly in medical content where precision is critical.

2026 - Present

Education

Work History