This is a work-in-progress project where I build a complete Transformer pipeline from the ground up. The goal is to implement one component at a time, in order: tokenizer → embeddings → positional encoding → attention → feed-forward → layer norm → encoder block → full model → loss → training → evaluation.
Tanvir Ahmed Sijan
Prospective PhD Applicant
My research interests lie at the intersection of NLP and AI safety, focusing on human-AI alignment through human-centered evaluation, alignment methods, and LLM interpretability to develop safer and more trustworthy language models.