Jayden Steadman

I like to build and evaluate practical ML systems like retrieval-augmented generation pipelines, fine-tuned transformers and classical ML comparisons. I try to focus on measuring what actually works rather than assuming it does. Recent work includes fine-tuning DistilBERT for emotion classification, building local RAG pipelines with cross-encoder re-ranking, and running honest evaluations that document negative results alongside the wins. Interested in the practical, deployable side of AI/ML.