About Karibu
Karibu is a clinical AI company in San Francisco. We work on a problem that sits underneath almost every other problem in a hospital: the instructions clinicians are told to follow, and whether those instructions agree with each other.
Every service line writes its own rulebook. How to run each procedure, what readings are safe, when to stop and escalate. Those pages are written by different people at different times, and they drift, from each other and from the regulations the service line is inspected against. A clinician mid-shift is the person who meets that drift, usually alone, often at night.
Karibu reads the rulebook against itself and against the sources the service line registered, and returns where those sources disagree. Every finding quotes both sides and names who registered each source. A clinician decides which one governs. Approved policy then answers questions at the bedside, in plain language, on the phones staff already carry on shift.
What we believe
The decision belongs to the clinician.
We find the disagreement and put both sides in front of the person who is accountable for the patient. Karibu never rules on clinical content, and that boundary is why a health system lets us near their policies at all.
Precision over reach.
A finding that wastes a clinician's time costs more than the finding was worth, so we would rather return less and be worth reading.
The work is judged at the bedside.
Everything we build is used by someone standing in a unit, mid-shift, with a patient waiting. That is the only test that counts.
Current Openings
Two founding engineers.
San Francisco. Starting November.
We are coming out of our first clinical pilot with a commitment to expand, and scaling from there. You would be the first two engineers. There is no hierarchy between the roles.
Founding Engineer, Policy Intelligence
You own the checking side.
Document ingestion, retrieval, and the rule engine that reads a large set of policy documents and returns where the sources disagree. The work is LLM systems and evaluation, on messy real-world documents where the right answer is genuinely contested and a false positive costs a clinician's trust.
You will probably have built retrieval or reasoning systems that had to be right, and you will care about evaluation as much as about features.
Founding Product Engineer, Point of Care
You own what a clinician meets at the bedside.
The question-answering surface, and the learning content generated from approved policy. Full-stack, React and Python, strong interface judgement, shipping into a live clinical environment with real users on shift.
You will sit in sessions with clinicians and watch them use what you built. Most engineers either love that or find it unbearable.
Interested? nathalie@karibu.ai (learn more)
What we offer
You will set the engineering culture
Competitive base and meaningful founding equity
Compensation reflects that you are the first two engineers, not the twentieth.
San Francisco, in person
The team works together in the same room, especially at this stage.
You set the engineering culture
There is no hierarchy between the roles, and no culture yet to inherit. You write it.
How to apply
Tell us what you have built.
nathalie@karibu.ai(learn more)
Looking for the product instead? See the evidence behind Karibu
Copyright © 2026 Karibu Connect Inc. | info@karibu.ai | San Francisco, CA
Last updated: September 14, 2026.