Building the tools and capabilities for controllable biology.
Human biology is a system that changes every hour and is read once a year. Labs exists to close that gap: to measure the body continuously, estimate its state, act on it, and learn from what the body does next.
01 · Measure
02 · Infer
03 · Act
04 · Learn
Biology becomes controllable the moment it becomes observable, interpretable and answerable — in that order.
The stack we are building
Four capabilities. One loop.
01
Measure
Multi-omic depth on a schedule — DNA, microbiome, blood, body composition, medical history — plus daily and instant signals from the same person, on one clock.
Periodic omics
Continuous daily signal
Clinical record ingestion
Signal quality and noise models
02
Infer
The body's state is never observed directly. We estimate it: a live latent model per person that updates with every new piece of evidence and carries its own uncertainty.
Per-patient state estimation
Drift and trajectory detection
Uncertainty as a first-class output
Cohort-level priors
03
Act
An estimate that does not change anything is a report. Labs builds the intervention layer: what to change, at what dose, in what order, and how early.
Intervention policies
Dose and sequence logic
Timing against predicted damage
Clinician-in-the-loop controls
04
Learn
Every measured response re-enters the model. The trajectory — intervention attached, response attached — is the training data that does not exist anywhere else.
Response labelling
Trajectory datasets
Asset-specific biological models
Continuous evaluation
Each capability feeds the next, and the last one feeds the first. That loop is what makes biology steerable instead of merely described.
Programs
What Labs works on.
Biological state models
Per-patient latent models that stay current between clinical visits, built to be read by both a clinician and a molecule designer.
Response and non-response
Why the same molecule at the same dose produces opposite outcomes — and detecting it while there is still time to act.
Intervention engines
Nutrition, supplementation, microbiome and behaviour as controllable inputs with measurable, attributable effects.
Trajectory data infrastructure
Consent, provenance, governance and the pipelines that turn treated patients into research-grade longitudinal records.
Clinical evidence
Studies run inside the live system rather than beside it, so validation and product improvement share the same loop.
Pharma-facing models
Asset-specific cohorts and biological read-outs for partners who need to see their molecule acting in living humans.
How we work
Principles.
Nothing ships that cannot be measured.
Every capability has to produce a signal we can read back later, or it does not go in.
Uncertainty is stated, not hidden.
A model that cannot say how sure it is cannot be used to change a therapy.
Control before prediction.
Prediction only counts when it arrives early enough to change the trajectory.
Built inside a running system.
Labs works on patients who are already being helped today, not on archived files.
The largest biological model is lived by every single of us, one human trajectory at a time.