Gero
GeroHacking aging
with physics and AI
The Science

Aging, read as a physical process.

We use the tools of physics and data science to describe how human health changes over a lifetime, then build AI that acts on that description. This page lays out the thesis, the framework, and the evidence behind it.

Thesis

Treat the organism as a complex system, not a list of broken parts.

Conventional biology studies aging one pathway at a time. Gero studies the dynamics of the whole system: how an organism's state moves through time, where it loses resilience, and how quickly it recovers from stress. Resilience itself becomes a measurable quantity.

That shift matters because it points to control. If aging is a drift in a dynamical system, then the right intervention is one that changes the system's dynamics, not one that patches a single symptom.

Dynamical model of aging

A framework in three moves.

How the physics-first view becomes something you can test, model, and build on.

01 · Measure
Quantify resilience
From large biomedical datasets we extract markers of how well an organism resists and recovers from stress. Loss of resilience is the signal that tracks biological aging.
02 · Model
Map the dynamics
We fit physical models to those trajectories to find the control points, the variables where a change in the system has the largest downstream effect on aging.
03 · Intervene
Design therapies
Generative AI turns those control points into candidate molecules and targets, which feed directly into the discovery platform.
Evidence

Grounded in the published record.

The founding science is documented across more than 75 peer-reviewed publications. The work below represents the direction of the program.

  • Dynamics of aging

    Work modeling the loss of resilience and the divergence of biological state with age, framing aging as a transition in a dynamical system.

  • Generative chemistry

    ProtoBind-Diff: target-conditioned ligand generation from protein sequence using a diffusion model. Released openly for the research community.

  • Benchmarking

    Harvest: an open dataset and benchmark for evaluating discovery models on biologically meaningful tasks.

Full publication list and preprints available on request via info@gero.ai.

From science to platform

Where the research becomes a tool.

The framework above is not academic for its own sake. It feeds two open tools that anyone in the field can use and evaluate.

Collaborators

Built with leaders in longevity and discovery.

Prof. Brian Kennedy Distinguished Professor, National University of Singapore. Director, Centre for Healthy Longevity, NUHS. Former President & CEO, Buck Institute.
Connected Discovery Founder and COO, ex SVP of Drug Discovery at BenevolentAI.
Medicinal chemistry Co-inventor of XELJANZ and CIBINQO; chemistry leadership from Pfizer, Eli Lilly, and WuXi AppTec.
Academic partners Associate Professor, National University of Singapore, and adjunct roles across leading aging research institutions.
Meet the full team