Your AI Scored 100%. Half the Evaluation Is Missing.
Audit missing model-evaluation grades before trusting a perfect score. Reproduce a CSV coverage report with 11 tests and explicit denominators.
BY JASON EAD
AI engineering & business applications
Original guides and analysis about coding tools, models, GitHub projects, agents, and creative AI workflows.
Filtered articles
6 articles
Your AI Needs Sources. Start With Search You Can Inspect.
Build a local SQLite FTS5 procedure-search baseline. Reproduce useful keyword hits and a revealing synonym miss before comparing more complex retrieval.
Your Cheapest AI Model Might Be Your Most Expensive Workflow
Include human review time when comparing AI workflows. Use a tested decimal calculator and a task-trial record, with clearly fictional inputs and no model-performance claims.
Search Business Procedures Locally With FastEmbed
Run a local embedding model over business procedures, return traceable sources and test the questions your library cannot answer. Includes complete Python code.
Catch AI Routing Regressions That Accuracy Alone Misses
Build a tested Python scorer for AI ticket routing. Use fictional cases, severity weights and strict run checks to catch regressions hidden by overall accuracy.
Validate AI Output with Pydantic Before It Reaches Your CRM
Build a tested validation boundary for AI-generated business data with Pydantic, strict types, source checks and a separate human-review step.