
Luminid exists to give people a fairer chance at work.
I'm Gabriel Golfin Araya. I founded Luminid in 2026 after seven years of recruiting at companies ranging from Amazon and Grupo Bimbo to firms in the legal services space. Over those years, I watched the same pattern play out across industries. Candidates with the right skills filtered out for the wrong reasons. Resumes that told one story when the actual capability told another. Decisions made in seconds based on signals that didn't predict the work.
The research had been clear for decades. Work samples predict job performance dramatically better than resume screening, and yet the entire applicant tracking software industry was built around the signal that worked the worst. The gap between what the field knows and what the software does is the reason Luminid exists.
I started Luminid to close that gap. Not by adding skills-based hiring as a feature alongside resume screening, but by building a platform where demonstrated capability is the primary signal and resumes are not scored at all. The methodology is published at /methodology. The architectural reasoning is in the blog. The work continues.
Luminid is built in Costa Rica. Latin America is where I'm from, where I've done much of my recruiting work, and where the hiring software market is most underserved by the incumbents. Luminid starts with candidates. You prove real skills through short work simulations, general business tasks like handling a hard client conversation, working through a decision with data, or closing a deal, and you build a verified record you can take anywhere. The company side comes later. For now the value is simple. Do the work once, and carry the proof with you.
Building from Costa Rica has shaped how Luminid thinks about hiring. We're focused on candidates and recruiters whose workflows do not fit the assumptions baked into incumbent US enterprise HR tech.
Luminid is in early development. The product is being built deliberately rather than rushed to market. The methodology is published before we have a long customer track record because methodology should be public from day one, not after we've accumulated reasons to defend specific outcomes.
Luminid is being built sustainably rather than on milestone-driven funding cycles. The pace is set by what's right for the product and the methodology, not by external timelines.
The work happens in the open. Material changes to the product and methodology are tracked in the changelog at /changelog. Decisions about how Luminid handles AI scoring, bias, and candidate data are documented at /methodology.
All inquiries go to hello@luminid.org. General questions, press, partnership, research collaboration. Inquiries are read and answered by Gabriel directly.
Candidates can write to the same address with questions about how verification works, what a verified skill means, or how the methodology applies to them. We read and answer directly.
Luminid is in candidate-first early access. If you have been filtered out of roles you could do well, the platform exists for you. Complete a simulation, earn a verified skill record, and take it with you anywhere.
The company side of Luminid, where recruiters and hiring managers evaluate candidates on demonstrated capability rather than resume proxies, is in active development and not yet open. Teams who want to be first when it launches can join the waitlist.
The methodology is at /methodology. The first blog post is at /blog. The work continues.