The Measurement Stack
TimeTrace Labs is a mathematics and artificial intelligence company building the measurement infrastructure of healthcare.
The early signatures of disease are continuous, subtle, and buried in biological noise long before they reach a clinic. The bottleneck in healthcare AI is not the model. It is the measurement underneath: sparse, irregularly sampled, multimodal, with noise that varies by sensor and subject, and with missingness that often carries information about the underlying biology. The standard tools applied to this data were not designed for its structure. Bigger models alone do not close that gap. Biology measured under constraint yields its structure more honestly than biology averaged over scale.
We build from mathematics upwards.
This is a deliberate choice. The mathematics is not simply preprocessing. It is the load bearing infrastructure of everything above it. At the base: stochastic processes, statistical inference, dynamical systems, the geometry of measurement under constraint. Higher in the stack: regularisers that enforce the structure a physiological signal must obey, representations that separate what persists in the body from what is artefact, and world models trained over multimodal time series in collaboration with NVIDIA. Between and throughout: the components that handle informative missingness, calibration, noise characterisation, the recovery of meaningful change from irregular and fragmented sampling. Each of these is a foundational problem. Solved or improved upon, each unlocks a large number of downstream consequences across modalities, indications and scales.
Most companies concentrate on one layer. We own the full vertical stack, coupled into a single platform. Some datasets are too small or too sparse for deep learning and the mathematics carries the work alone. Others have the depth to support world models on top. Many require both in sequence. Because we own the entire stack, we meet the data where it is rather than forcing it into a shape that suits a single architecture.
In this series we will share components of the stack and explain how each connects to the platform as a whole. All modules are integrated. The stack is designed to evolve with the field: to absorb new architectures, new mathematics, and the expanding ability to build, discover and validate more from less. We are at the beginning of an era of continuous and compounding discovery in how the body is measured. We are building the measurement infrastructure for this era.
Please cite this work as:
TimeTrace Labs. The Measurement Stack. TimeTrace Labs Blog, April 2026.
BibTeX:
@article{timetrace2026measurementstack,
author = {{TimeTrace Labs}},
title = {The Measurement Stack},
journal = {TimeTrace Labs Blog},
year = {2026},
url = {https://www.timetracelabs.com/blog/timetrace}
}