Secure the software the world runs on.
Pensar is a small team of operators, researchers, and engineers building adversarial AI agents that continuously attack production code — so the next exploit hits us before it hits the banks, hospitals, and infrastructure providers we work with. We do it on the frontier of offensive security, ship Apex in the open, and care a lot more about what gets delivered than how many meetings it took to get there.
What makes the work different — and worth doing.
Mission first
Our customers run banks, hospitals, payments rails, and the infrastructure underneath them. The job is to find the exploit before the people on the other side do — and to keep finding it, every day, as the systems grow. That outranks every roadmap argument we'll ever have.
Adversarial by default
If a system claims to be secure, we assume it isn't. Every product decision is run against the attack it has to survive. Every release ships only after our own agents try, and fail, to break it.
Frontier-grade research
Offensive AI is a moving target. We hire researchers who push the field forward, pay attention to where the bleeding edge is going, and build the tooling that lets the rest of the industry catch up to it.
Open in the open
Apex is public. Our research, advisories, and CVEs are public. Pensar's credibility comes from what we put on GitHub, not what we put behind a paywall — and the work you do here gets your name on it.
1 role open in New York.
What you get for shipping with us.
Don't see your role? Send us what you've built.
The best people we've hired didn't come in through a job posting. If you've shipped offensive tooling, broken an interesting system at scale, or want to spend the next stretch of your career securing the software the world runs on — send us a note.