Geofrey Mwangi

I am an AI evaluation specialist focused on testing and optimizing LLMs and autonomous systems to ensure safety, reliability, and real-world readiness. I lead red teaming exercises, design rigorous tests, and dig into data to validate model behavior.

I have hands-on experience in AI training tasks including data annotation and labeling for text, image, and audio with high accuracy, and I write evaluation scripts in Python using tools such as Hugging Face Evaluate, LangChain, and SQL for deep data analysis. I am skilled with Labelbox, Prodigy, Amazon Mechanical Turk, and GCP for dataset workflows and management.

Detail-oriented and highly collaborative, I combine technical QA skills with strong communication and cultural adaptability to help teams deploy AI people can trust. I also have domain strengths in medical transcription, research writing, and reporting, and I write about AI advancements in English and German.