Things that should exist. Things that should be autonomous. Hardware and software, built end to end.
Some things don't exist yet. Some things aren't autonomous yet. Both are reasons to build.
Hardware and software, built end to end under one roof.
A sub-kilo autonomous drone for close-terrain survey, inspection, and search. Built in-house because the drones that fly themselves, instead of asking you to, didn’t exist yet.
Request the spec →Force-sensing, compliant actuators for humanoid and quadruped joints. Built in-house because robots that share space with people need joints that can feel, not just move.
Request the spec →A macOS dictation app. Hold a key, speak, your words land where you’re typing. Runs entirely on your Mac because privacy belongs in the architecture, not in a settings panel.
Visit openear.fyi →Our own chatbot, in the same family as Claude and ChatGPT. Built because a tool we use every hour should be one we have full taste over.
Try Mally →An open-source virtualization platform. A copyleft alternative to VMware. Every fork stays open. Self-hosted. Next on the bench.
Follow progress →An AI co-pilot for Open VM. Setting up, tuning, and debugging a VM platform has historically required a senior infrastructure engineer. Vantage replaces that ask with an agent you can talk to.
Follow progress →We build across layers of the stack — on-device, product, infrastructure, and the AI co-pilots that make infrastructure easy. Different shapes, same small group.
Pick the right abstraction.
Build it ourselves.
Ship to real users.
Mechanical, electrical, firmware, and software under the same roof. Fewer hand-offs, shorter feedback loops.
We build things that fly, turn, and run. Papers, decks, and demos come after — if at all.
On-device when privacy is the product. Our own chatbot when taste matters. Open source when an ecosystem should own the code.
We pick a short list of problems and go deep. Everything else we say no to.
Building something
autonomous?
We’re hiring engineers across mechanical, firmware, and ML — and we’re open to research and production partnerships.