Comparing prompt engineering, RAG (LangChain + ChromaDB), and LoRA fine-tuning for persona-consistent LLM chatbots. Built with LLaMA 3.1 8B, Gradio, and HuggingFace.

Unsupervised clustering on synthetic banking data to segment customers into financial persona types — covering spend behaviour, income, debt, and demographic profile.

Source for the Experience Notation documentation site. Includes language spec, tutorials, and examples.

Experience Notation (.expn) is a human-readable DSL for modelling structured user journeys. It bridges qualitative design with simulation and LLM-driven analysis.