About Passivia
Engineering intelligence for processes that cannot fail quietly.
Passivia works where an incorrect inference has a physical consequence: an embryo, a membrane train, a season of yield. That constraint shapes every engineering decision we make.
Operating principles
How we build.
Determinism over novelty
A system that behaves predictably at the ninety-ninth percentile is worth more than one that impresses on average. Latency, failure modes, and handback behaviour are specified before models are trained.
Evidence as a deliverable
Validation reports, provenance records, and audit trails ship with the product. Regulatory readiness is a design input, not a later phase.
Human authority preserved
Clinicians, plant operators, and agronomists retain override authority at every decision point. Automation removes labour, not accountability.
Edge-first by default
Inference runs next to the process. Data residency, connectivity loss, and latency budgets are constraints we design for, not exceptions we apologise for.
Trajectory
From a vision laboratory to four deployed systems.
Foundations
Embedded vision research applied to time-lapse embryo assessment.
Clinical
IVF Watch validated across multi-centre retrospective cohorts.
Robotics
AutoICSI micro-manipulation platform enters pre-clinical validation.
Industrial
DesalPure deployed in coastal reverse-osmosis pilot operations.
Agriculture
AgriSense scaled to permanent crops across three continents.
Work with us