Builders, not bystanders
Everyone here ships. We hire people who would rather make the thing than manage the people making the thing.
We're a small team building private AI for personal hardware — engineers, researchers, and product builders who care about privacy by default. The work is hard, the ownership is real, and the bar is high. If that sounds like you, we should talk.
“We're building AI that expands what people can do on their own terms. That only happens when the team cares about the last 10% as much as the first 90%.”
You get meaningful responsibility from the start. Our teammates have shipped systems software, compiled models for real hardware, and launched products people use every day. In the same week, your work might touch the runtime, the compiler, and the product experience.
A few principles we actually hire and operate by.
Everyone here ships. We hire people who would rather make the thing than manage the people making the thing.
Powerful AI should not require sending your work somewhere else. We sweat the details so private inference feels practical, not precious.
People run models on files, code, and workflows that matter. Correctness, privacy, and judgment are non-negotiable, not features we get to later.
Small teams, short feedback loops, real ownership. We'd rather learn on real hardware than debate in a doc.
Strong opinions, held loosely. The best idea wins, regardless of who it came from or how senior they are.
You're trusted with too much and bored by nothing. We hand you the problem and the latitude to solve it.
We build for real laptops and desktops, not benchmark fantasies. The best solution runs well on the machine someone already owns.
Local-first is not a feature toggle. It is how the product is designed, not something users opt into after reading the fine print.
Compiler, runtime, and product are one problem. We optimize across the stack instead of polishing one layer in isolation.
Everyone here ships. We hire people who would rather make the thing than manage the people making the thing.
Powerful AI should not require sending your work somewhere else. We sweat the details so private inference feels practical, not precious.
People run models on files, code, and workflows that matter. Correctness, privacy, and judgment are non-negotiable, not features we get to later.
Small teams, short feedback loops, real ownership. We'd rather learn on real hardware than debate in a doc.
Strong opinions, held loosely. The best idea wins, regardless of who it came from or how senior they are.
You're trusted with too much and bored by nothing. We hand you the problem and the latitude to solve it.
We build for real laptops and desktops, not benchmark fantasies. The best solution runs well on the machine someone already owns.
Local-first is not a feature toggle. It is how the product is designed, not something users opt into after reading the fine print.
Compiler, runtime, and product are one problem. We optimize across the stack instead of polishing one layer in isolation.
We invest in benefits that help you do your best work and live well outside of it.
Comprehensive medical, dental, and vision plans so you and your family are covered.
Paid parental leave so you can be present for the moments that matter at home.
Take the time you need to rest, recharge, and show up ready to do great work.
Competitive salary and meaningful equity so you share in what we build together.
A learning and conference budget to keep your skills sharp and your curiosity funded.
A monthly wellness stipend for gym memberships, therapy, fitness, or whatever keeps you balanced.
Equipment and setup support so your workspace at home is comfortable and ready to go.
An annual team recharge week to step away, reconnect, and come back with fresh energy.
Find where you fit. Don't see your exact role? Reach out anyway.
We're always looking for ambitious people. Make a role for yourself by emailing careers@orafrontier.com.