Joel Waweru Njoroge User

I’m an AI Training Specialist and AI Data Specialist and Trainer focused on preparing, labeling, and curating clean, high-quality datasets that are ready for machine learning and artificial intelligence model training. I bring a detail-oriented approach to make sure data is accurate, consistent, and usable for performance gains.

In my freelance work, I annotate datasets across images, text, audio, and rubrics to improve AI accuracy for speech recognition and computer vision. I also perform audio transcription and validation—converting speech to text for training natural language processing models with strong attention to quality.

I go beyond raw labeling by generating custom prompts and responses for AI-driven chatbots and virtual assistants. I’ve also built and applied structured evaluation rubrics, writing detailed evaluations to support reliable, objective model improvement.

My results reflect that craftsmanship: I achieved 95% accuracy in data annotation, increased transcription speed by 20% while maintaining 98% accuracy, generated 50+ customized prompts improving chatbot response accuracy by 30%, and labeled 10,000+ handwritten samples to enhance handwriting recognition accuracy by 25%. I’m motivated to keep delivering dependable training data and measurable improvements as an ongoing training partner.