Build the future people actually own.
Personal Supercomputing - AI agents that run on your devices, not in someone else's cloud.
your agents, your computers, your information, your intelligence - really private and personal and yours forever
Builders wanted.
We are going to give every person on Earth their own intelligence. Not a chatbot in someone else's cloud - a supercomputer of your own, agents of your own, memory of your own, on hardware you own, under a consent protocol that makes your rights enforceable in code.
The other future is already under construction. Every year, more of your life is computed by machines you don't control, under incentives you didn't choose. We think that future is unacceptable - and beatable. So we're building the whole stack to beat it: the protocol (PCHP), the platform (Agent One), the hardware (Puppy One), the marketplace (the 🤫 Yellow Pages) - in the open, where you can watch us do it.
Who are we hiring? People who have shipped something they love and can prove they use it. Systems thinkers who take responsibility for outcomes, not tasks. Engineers who read the spec and rewrite it shorter. Sellers who can walk into the world's largest institutions with a straight face and a working demo. People who find "that's impossible on-device" to be a personally offensive statement.
The deal is honest: the problems are brutally hard, the standard is Apple-grade, the team is small, and your name goes on what ships. If your personal data is the next asset class - and it is - somebody has to build the vault, the rails, and the keys. Come do the best work of your life.
- Manish, Founder 🤫
Why hussh exists
In 2021, our founder's identity was stolen. Bank accounts, Social Security number, driver's license - taken, sold, used. Nobody told him.
That day left one rule:
If anyone touches your personal data, you know - and nothing moves without your consent.
Everything we build follows from that sentence. We've written up the case for making that right enforceable - read it before your interview and you'll be ahead of most.
What we're building
A consent protocol with cryptographic audit trails. On-device intelligence under real latency and memory budgets. Local-first data with encrypted, scoped export. Privacy as an invariant checked in code - not a promise made in copy.
Agent One
The personal agent platform. A private home for your data and the agents that serve you.
See One →Kai
The first specialist: your money. A family office loyal to exactly one person.
Kai for advisors →Nav
The guardian. Inspects every request that touches your data.
Privacy & ownership →PCHP
The consent handshake protocol. Every access visible, auditable, revocable.
Read the protocol →All of it ships on Puppy One - the computer that does the work. Protocol under platform, platform under agents, agents under you. The human is the root of the system.
How we work
You own your work - a product surface, a metric, a customer relationship - and it ships with your name on it.
The standard: correct, then clear, then secure, then fast - in that order. If two hundred lines can be fifty, they become fifty. If a feature needs a manual, the feature failed.
And one test for every decision: does this put the person in control of their own intelligence and data? The long version lives at how we work, and the receipts live at building in the open.
Open roles
129 roles are open right now, across nine cities. Every one of them takes an application on its own page, in about a minute.
Head of the 🤫 Academy (Train-the-Trainer Lead, L&D)
A funded US FTE to build our Learning & Training Academy: a train-the-trainer L&D engine that teaches agentic AI and puts 🤫 Agent One in everyone's hands.
Chief Revenue Officer (CRO)
Own revenue and adoption end to end. Land and grow the world's largest companies and governments.
Full-Stack Engineer (AI · systems · data · UX-first)
Rockstar full-stack engineers who are AI-, systems-, data-, and user-experience-first - all at once.
Mobile Engineers (iOS & Android) - the app people live in
The best mobile engineers alive, building the 🤫 Agent One app people open every day.
Spatial & Wearables Engineers (visionOS, glasses, watch, VR)
The best glasses, watch, and headset builders - Agent One on your face and your wrist.
Forward-Deployed Engineers (whole team)
Build it with the customer, on their terms - a whole team embedded where the work and the data live.
AI / ML Researcher-Engineer (One)
Hardcore AI/ML researchers and engineers with the curiosity to build great AI, data, and systems.
Every level, every geography - search all 129 by city and team. Don't see your role? Apply to the closest one and make the case in your own words - the form asks.
How hiring works
- 1
Apply
The form on the role's own page. Four boxes and your resume, about a minute. You get a reference number straight away, and a page of your own to check where things stand.
- 2
Sign the NDA
Before any interview, a simple mutual NDA. You keep a copy.
- 3
Build
One week, any stack: a personal data agent that runs on your own real data. If you still use it a week later, it worked. Start from our developer surface or from scratch.
- 4
Talk
The team reviews your build and sends feedback. Then one 15-minute conversation.
- 5
Decide
A real answer in days, either way.
🤫 Your agents. Yours to own.
© 2026 HushOne Inc. · Kirkland, WA · An equal opportunity employer.