A solo maker used Codex to collect 7.1 million job listings
In a solo-build example, was used to take a list of company names, find each company’s careers page, and collect its open jobs. The goal was to avoid job boards like LinkedIn and Indeed, where many listings can be stale, indirect, or posted by outside agencies. Large-scale was difficult before because every company careers page can look different.
helped handle that messy page-by-page work. then turned the collected text into JSON with fields such as salary and years of experience. The method produced 7.1 million job listings, including more than 220,000 .
The data became a public service called Hiring.Cafe, with filters for job titles, job , excluded terms, industries, individual contributor versus management roles, and experience level.
Key points
- was used to find company careers pages and collect job listings from them.
- converted the collected job text into JSON fields such as salary and experience level.
- The reached 7.1 million job listings, with over 220,000 .
- Hiring.Cafe exposes filters for role, industry, management level, exclusions, and experience.
- The main maker lesson is that AI tools can help build large data products, not only write app code.