Imaobong udoukpong

Data Annotator and Prompt Generator for LLMs

Technology / Internet

Calabar, Calabar Municipal

Services offered

Data Annotator that specializes in:

Annotation and evaluation of large-scale datasets used to train and validate AI systems across healthcare, social media, and information retrieval domains.
Performing LLM response comparisons, assessing accuracy, coherence, safety, and alignment with clinical reasoning standards.
Conduct audio transcription and segmentation, ensuring linguistic accuracy, contextual correctness, and consistent labeling.
Evaluate and annotated images and videos for object recognition, visual changes, and scene understanding relevant to AI training.
Apply strict, rule-based annotation guidelines with clinical-level precision and thorough documentation.
Contribute to datasets requiring contextual medical understanding, patient safety awareness, and ethical judgment.

Video annotation

Audio transcription

Image annotation

Llm prompt generation

Data annotator

Approximate rate: USD$ 10 per hour

Experience

Data Annotator and Prompt Executor at Appen (June 2023–October 2025, remote):

Annotated large-scale datasets for AI training in healthcare, social media, and information retrieval, including LLM response evaluations, audio transcription, and image/video analysis with clinical precision. Contributed to numerous tasks, improving AI model performance by spotting errors and edge cases while exceeding productivity benchmarks in international teams. Also specialized in video annotation to detect visual/temporal changes for healthcare, social media, and moderation projects. Made sure to refine annotation quality through iterative feedback, enhancing downstream AI accuracy.

Experienced in prompt execution tasks at Appen, crafting targeted inputs for generative AI models like ChatGPT to test safety, coherence, and edge cases in healthcare and social media domains. Applied advanced writing skills to generate original, rule-compliant prompts with hateful or adversarial content, improving model robustness against harmful outputs. Demonstrated native-level English proficiency, attention to detail, and creativity to meet complex guidelines, contributing to high-quality datasets for LLM fine-tuning.