I help AI teams fine-tune language models on complex, culturally rich subject matter that most training pipelines get wrong. My work spans structured Q&A dataset creation, model output auditing, and prompt engineering — with particular depth in African spiritual philosophy and Yoruba cosmology (the Ifá corpus), where Western-trained models frequently produce biased, flattened, or factually inaccurate responses.

If your model needs to represent underrepresented cultures, languages, or knowledge systems accurately — not just fluently — this is the gap I close.

Build structured, source-grounded Q&A datasets for LLM fine-tuning

Audit model outputs to catch cultural bias, factual errors, and low-resource language translation issues

Design precision prompts and system instructions for consistent, accurate outputs on nuanced subject matter

Review and correct hallucinated or culturally inaccurate AI-generated content