Filtering Millions of H-1B Records to Find Real Hiring Signals

Funding databases tell you which companies have money. They don't tell you which companies are actually hiring. That gap is exactly what a new data source closes: if a company files for H-1B sponsorship, that filing goes straight into a government database, and it turns out to be a far better hiring signal than who raised a round.

Why sponsorship beats funding as a signal

The insight here came from a simple observation: when a company sponsors an H-1B visa, it has to go through a formal government process, and that process creates a public paper trail. Unlike startup and funding databases, which only tell you who has capital, sponsorship records tell you who has actually gone through the work of hiring someone and backing that hire with a federal filing. That makes it a higher-quality signal for identifying companies that are genuinely bringing people on, not just companies that raised money and might expand someday.

The scale of the data

These aren't small datasets. Beyond the startup and funding databases already being used, there are records from the Department of Labor, USCIS, and other federal sources covering every employer that has ever submitted an H-1B application. The numbers run into the millions, with sponsorship history data that looks to be in the range of six million-plus entries. That scale is exactly why the data can't be used raw. It needs to be filtered down to something a hiring-signal pipeline can actually work with.

What to filter for

Pulling in millions of records only helps if they can be narrowed to what matters. The plan is to check each filing for several things: the company's sponsorship history, the filing dates, the frequency of filings (a company that has filed once looks very different from one that has filed five hundred times), whether the company is still active, and how recent its most recent sponsorship activity is. Some of the companies in these datasets will be out of business by now, but since the records include filing dates, it becomes possible to filter down to recent, relevant activity rather than treating every historical filing as equally useful.

Where this fits in the pipeline

This data source is set to slot into the pipeline after the data is pulled in, joining the other sources already being collected: scraped websites, job pages, social links, company addresses, and more. It becomes one more high-quality layer in a broader system built to surface real hiring signals rather than guesses based on funding alone.

Key takeaways

• H-1B sponsorship filings are public government records, since sponsoring an employee legally requires going through a formal filing process.
• Federal sources like the Department of Labor and USCIS hold sponsorship data for every employer that has ever filed, running into the millions of records.
• Sponsorship history is a stronger hiring signal than funding data because it reflects actual hiring activity, not just available capital.
• Filtering criteria include filing frequency, recency, and whether the company is still active, since not every historical filer is still operating.
• This data is meant to be combined with existing sources like scraped websites, job pages, and social links, not used as a standalone signal.

Who this is for

This is for anyone building lead generation or hiring-signal pipelines who wants a look at how public government data can supplement funding and scraping-based sources. It's a behind-the-scenes look at the ongoing 80 Days to Stay data project.