ACME Competitive Wage Intelligence

An executive dashboard answering one question: are we paying entry-level staff enough to compete with the employers down the road?

What it was

Built as a demo for a C-suite audience, it compared ACME's hourly wages at 148 yard locations against nearby competitors such as Walmart, Amazon, Home Depot, Costco and Starbucks, adjusted for local cost of living.

It also showed how to build an AI app where LLMs act as steps in a fixed pipeline (extract a wage, classify a job title, write a summary) rather than as autonomous agents. It used only free data: employer career sites, BEA regional price parities, and BLS state wage baselines.

How it worked

• Playwright scrapers collected live job postings from employer career sites.
• LLM steps extracted the pay from each posting and classified the role, each step on its own model via OpenRouter.
• Inverse-distance weighting matched each yard to the competitor postings around it; a what-if simulator showed the effect of a raise.
• APScheduler refreshed the data on a schedule, and an LLM wrote the national narrative at the top of the page.

Built with