Evaluation-Driven Development

A live demo of a lightweight n8n‑based framework that uses evaluation tests as acceptance criteria, regression guards, and guides metric‑driven fine‑tuning and cost optimization.

Overview

Live demo of a lightweight n8n-based eval framework
Using evals as acceptance criteria and as regression guards
Metric-driven fine-tuning and cost optimization. Comparing GPT-5, GPT-5-chat, Grok-4-fast-no-reasoning
Turning customer feedback into new evals

Tech stack