AI Trainer & Data Labeler (LLM training, dataset building, output review)
Trained and fine-tuned small language models to improve response relevance and quality for real-world use cases. Reviewed and corrected AI-generated outputs by providing feedback to reduce hallucinations in customer-support chatbot responses. Built and supported dataset creation efforts for local language processing to better serve African markets. • Used Label Scale AI and custom scripts to train and fine-tune LLMs. • Performed evaluation-oriented review of model outputs and iterated on improvements. • Collaborated with a startup team to assemble a dataset for local language processing. • Supported domain adaptation for African markets through dataset and training contributions.