How Jobright uses AI to help foreign workers navigate the US job market | TechCrunch

Job searching can be daunting, especially if you are a non-immigrant in the U.S. Not only do you have to compete on the merits of your resume, but you also need to scour through hundreds of postings to identify the few that sponsor work visas. And if you lose your job, you only have 60 days to land another position before you have to leave the country.

The lottery — the H-1B program that randomly selects skilled foreigners to work in the U.S. — is notoriously difficult to win. For the fiscal year of 2025, U.S. Citizenship and Immigration Services (USCIS) received nearly half a million entries, according to official data. About 120,000 of them were selected to move onto the “registration” phase, but eventually, only 85,000, or 18%, of all applicants will receive a visa.

A California-based startup called Jobright.ai wants to make job search easier for foreign workers in the U.S. As an immigrant, company co-founder Eric Cheng knows firsthand the struggles of foreign workers, so it was a no-brainer to build an “H1-B filter” into Jobright, he told TechCrunch. The feature, which is based on USCIS data of companies that have historically offered H1-B visa sponsorships, became an instant hit among job seekers from India and China.

“This group has long been overlooked,” Cheng said, who previously worked as Box’s early engineer. To date, around 30% of Jobright’s users are foreign workers. And thanks to its H-1B filter, Jobright has amassed 50,000 registered users since its launch in April, without spending on marketing.

During the COVID-19 pandemic, as he travelled across China after selling his content collaboration SaaS company, Cheng conceived the idea of a job-seeking platform. Over the months-long trip, he visited mostly underdeveloped cities, where he met dozens of young people struggling to land jobs amid a global economic slowdown. Cheng ended up helping 10 of them secure jobs simply by asking about their skillsets and suggesting career options.

“They were talented, but their understanding of the job market was pretty limited. They just didn’t know what jobs were out there and what they could possibly do,” Cheng said. “What if they all have a seasoned headhunter who knows them well and can recommend jobs that are beyond their knowledge? The arrival of ChatGPT gifted me the ability to build such a tool.”

Leveraging large language models, Jobright created an AI agent that acts as a headhunter tailored to individual job seekers. The AI interprets candidates’ experience and recommends jobs that they might not have otherwise considered. It scores positions based on their compatibility with the candidates, a feature used by 60% of Jobright’s users every day. It also helps users search for LinkedIn contacts who would likely offer them a job referral by identifying their alumni or former co-workers.

Despite competing in the crowded AI-powered job search market, Jobright thinks it has found a niche. While it shares some features — like using AI to write resumes and fill out applications — with more established players such as YC-backed Simplify, its target users are different.

“Most of our users are mid- to senior-level professionals with at least a couple of years of work experience,” Cheng said. “Their challenge is not applying for more jobs, but finding personalized job search strategies and opportunities that align with their career stage and professional strengths.” Simplify, on the other hand, is popular among internship-seeking students and fresh graduates, the founder observed.

“Yes, there are LinkedIn and other job listing platforms, but users still need to spend a lot of time searching on their own. The information asymmetry is huge,” he added.

When asked if he worried LinkedIn could develop similar AI features to replace Jobright, Cheng suggested that the giant is fundamentally a social network, whereas Jobright is designed to be “AI-native.”

“The cost of LinkedIn to fully embrace AI is high because it will have to topple its core ad-based business model, which is driven by clicks rather than accuracy. We aren’t just an AI wrapper. AI defines all aspects of our product, whether it’s data, recommendations or user experience.”

Jobright previously raised a $4.5 million round led by Lanchi Venture, followed by UpHonest Ventures and Source Code Capital. Down the road, the startup has plans to expand its user base outside the U.S.