AI Pulse · Under the Radar

A GitHub project called career-ops has passed 69,000 stars, which is a strong signal that a lot of people are looking for a more disciplined way to use AI in a job search. It turns tools such as Claude Code, Codex, or OpenCode into a personal job-search command centre. It can scan more than 100 company career portals using Playwright, score roles from A to H across five weighted dimensions, and draft an ATS-tailored CV for each promising listing.

The important detail is what it does not do. Career-ops describes itself as “a filter, not a spray-and-pray auto-applier”. It never submits applications automatically. That matters because the useful part of AI here is not sending more applications. It is sorting a noisy market, finding the better opportunities, and helping someone prepare a more relevant response without pretending there is no human judgement involved. The project has also already produced several forked or rebranded mirror repositories, which suggests the idea is spreading quickly. You can see the project details in the career-ops repository.

For business owners, the lesson goes beyond hiring or job hunting. Agent-style tools are becoming practical front ends for repeatable research and decision work. Give an AI clear criteria, access to defined sources, and boundaries on what it can act on. Let it rank, summarise, and prepare work. Keep final decisions and submissions with your team. This is the kind of thing we build into an AI command centre, where AI helps run a process without being left to run unchecked.