Intelligence Evolves. Matter Follows.

An evolving engine for molecular physics, trained within the polymer space and designed to reach every industry that depends on new materials.

Our platform Our focus

Many scientific questions have more possible answers than any lab could ever test. Our systems search those spaces the way evolution does, at the speed of computation, so only the most promising experiments reach the bench.

How the system works

The scientific question decides which data, models and experiments the system uses. Generative models and genetic algorithms do the search.

  1. Define the question

    Set the outcome that matters and the constraints that must hold.

  2. Choose the tools

    Bring in the data, models and experiments the question needs.

  3. Evolve

    Genetic algorithms select, recombine and mutate options against every objective at once.

  4. Learn and redirect

    Each result sets the next step, and the reasoning behind it is kept.

Why genetic algorithms

  • Many objectives at onceThey return a whole range of best trade-offs, not a single compromise.
  • No smooth problem requiredThey work on rugged, discrete design spaces where gradient methods stall, which is what real chemistry and biology look like.
  • Broad search before convergenceA diverse population keeps exploring after the first good answer appears.
GOAL A →GOAL B →Pareto front
Each dot is a possible solution scored on two competing goals. The highlighted points are the best trade-offs, where neither goal can improve without the other getting worse.

Designed so every result can be traced and reviewed

  • Evidence is recordedEach result keeps the data, models and assumptions behind it, with an estimate of how certain it is.
  • Scientists approve each experimentA scientist reviews the evidence and approves each experiment before it goes to the lab.
  • Failed attempts are keptRejected candidates and failed experiments stay on record, so later studies can see what was already tried.

Therapeutics first

The first application is drug delivery. Most long-acting medicines are adapted to a short list of existing polymers; Amphora designs a new polymer for each drug, using generative, physics-based and chemistry-rule models.

Linear
Star
Block copolymer
Crosslinked

  1. Design

    Generative models propose polymers against a full specification: release profile, safety, manufacturability and commercial viability.

  2. Predict

    Physics-based models rank candidates by how they will release the drug, before any material is synthesized.

  3. Learn

    Each batch is predicted before it is made, and the gap between prediction and measurement retrains the models.

Deployable across industriesThe same system is designed to extend from medicines to coatings and other materials fields, with data and models specific to each.

Our therapeutic starting point

Long-acting therapies, starting in the eye

Our therapeutic work begins in the eye. Approved drugs, small doses and fast lab readouts make it the clearest place to prove the engine.

Retinal disease

Wet AMD, diabetic macular edema and retinal vein occlusion are treated with injections into the eye every one to three months. A depot that releases the drug steadily could mean far fewer of them.

After eye surgery

After cataract and other eye surgery, patients put in steroid and anti-inflammatory drops several times a day for weeks, and many miss doses. A depot placed during surgery would deliver the full course on its own.

Dry eye in dogs

Dogs with dry eye need drops or ointment twice a day for life, which owners struggle to keep up. A long-acting depot would replace most of that routine.

Inherited retinal disease

Many inherited retinal diseases are caused by genes too large for the AAV viruses used in gene therapy today. Nonviral polymer carriers can hold much larger genetic cargo.

The name

An amphora is the vessel the ancient Mediterranean used to store and carry oil and wine.

We work at the molecular scale: an evolving engine that turns the physics of matter into new materials, for medicines first and then for any field that needs them.

Founders

Jeff Riley

Co-founder

Jeff Riley

Chief Executive Officer

Took three companies public; raised more than $1.5B for deep-science ventures.

LinkedIn
Alex Boches

Co-founder

Alex Boches

Chief Product Officer

Commercialized AI for scientific discovery. Multi-time biotech founder with an exit.

LinkedIn
Adam Breashears

Co-founder

Adam Breashears

Chief Technology Officer

25 years in AI and autonomous systems. Ran AI engineering at Metis.

LinkedIn

Scientific and development team

Ira Spector

Ira Spector, PhD

Head of Clinical Development

Wyeth and Allergan. More than 35 approved drugs.

LinkedIn
Ray Stapleton

Ray Stapleton, PhD

CMC Lead

20 years at Merck, including the Gardasil and Keytruda launches. Former CTO at Vaxart and Genocea.

LinkedIn
Jepson Taylor

Jepson Taylor

AI R&D

Co-founded Zeff.ai (acquired by DataRobot).

LinkedIn
Dean Farmer

Dean Farmer, PhD

IP Lead

Materials and formulation patent estates.

LinkedIn
Wes Spence

Wes Spence

VP Partnerships

25 years closing partnership deals.

LinkedIn

Advisors: Stephen Gately (TD2, CRO) and Yu Ge, PhD (formulation, medicinal chemistry, DMPK).

We partner with companies and researchers working on new polymer materials, in medicine and beyond.

info@amphora.bio