@santifer built a multi-agent job search system on Claude Code, ran 631 assessments, sent out 68 applications, and landed the Head of Applied AI role, then open sourced it.
The problem: Searching for senior AI roles is a full-time job. Read JD, map your skills, rewrite CV, fill 15 field forms, multiply by 10 offers a day. 74% of offers are a poor fit. You will know after reading 800 words. Solution: Carrier-Ops automates analysis. You make every last call.
What stands out?:
🎯 A–F scoring in 10 dimensions with role match gate filters
📄 ATS PDFs with per-roll keywords, reordered bullets, auto-region
🔍 Scanners at 45+ companies like Ashby, Lever, Wellfound
⚡ Batch mode runs 122 URLs in parallel with the worker architecture.
🧠 StoryBank generates reusable STAR+R interview responses.
💰 Negotiation script for Salary, Geo Gaps, Competitive Offers
📊 Dashboard with filters, sorting, and slow preview
🔁 680 URLs trimmed with human review before submission.
From the numbers:
631 reviews, 516 unique offers.
354 PDFs generated.
68 applications were sent.
9.1K GitHub stars, 1.6K forks
Different because it’s not an auto-apply bot, it’s a filter. The design principle is clear: automated analysis, not decisions. And the system itself is a portfolio, building a multi-agent architecture for multi-agent roles is the most direct proof of competence.
Perfect for senior engineers, AI practitioners, and anyone running a serious, high-signal job search at large.
Be a victim of the latest and greatest launches in tech, SaaS and AI at PS, follow to stay informed → @rohan suggests.