Applied technology lab
Carbon Based is an applied research lab. We put the two great breakthroughs of our age — machine intelligence and robotics — to work expanding, improving, and augmenting human health.
What we do
Three pillars
Pillar 01 · LLMs
A large model's ability to reason, infer and statistically determine becomes a platform for advanced research — hypotheses explored in silico long before the physical world is ever touched.
Pillar 02 · Multimodal
Multimodality unlocked the ingestion of real sensor data — vision, sound, signal — and the power to corroborate it with reasoning. The system perceives and thinks at once.
Pillar 03 · Robotics
Precise, repeatable motion gives sensors an outstanding signal-to-noise ratio; driven by our inference and multimodality, it unlocks use cases neither could reach alone — and paves the way to intervening in the medical process itself.
In the lab
Project 01 · Pellicle
A fine-tuned model that synthesizes human skin and predicts how it reacts with compounds — so a formulation can be tested in silico long before it is ever mixed.
$ simulate --pair "GHK-Cu + retinaldehyde" --layer epidermis
binding affinity0.82
penetration depth118 µm
irritation index0.14 LOW
predicted synergy+23%
degradation · 24h6.1%
status: stable · confidence 0.91
$
Project 02 · OmniVision
Vision- and inference-powered AI that looks after the elderly at home or in care — posture, falls, presence, intruders, recognition — from cameras alone, processed on-device.
Project 03 · OmniRobotica
A collaborative robotic arm with swappable sensor heads — ultra-precise, non-invasive health monitoring, hands-free. Work in progress, early-stage research.
Why we exist
Inference, multimodality, and hardware are how we get there.