Better Decisionsevery Experiment

Predicting biological and chemical processes. Built for the infrastructure you already have.

Start here

Pharma & BiotechLab-in-the-loop assay and process development: find the right parameters automatically, use data from past runs and characterise sources of variation.For example:LNP formulationiPSC differentiationCHO culturesDownstream purification
AI LabsPhysically execute and validate your model’s predictions, fully automated, on your own infrastructure or that of our partners.Each round of prediction, experiment and measurement yields ground truth validation data, which makes the loop a natural environment for post-training.

Every experiment or assay follows a protocol. Every setting in it is a choice.

Ten settings, three values each: 3 to the power of 10 = 59,049 possible protocols.
Reproducibility is notoriously hard to get right. What works in other labs may not work for your cells.

You can’t test them all, but you can learn which parameters matter most.

ProviGen helps you get to your target faster. It’s an active learning model recommending each next experiment for maximal understanding, so your process reaches its target in fewer runs.

Every result steers the next experiment.

Select a step to hold the figure there.

Pick the protocol parameters worth running.
Each design automatically becomes an executable protocol variant.
Run the protocol on your instruments with selected parameters.
Results are automatically acquired and compared to your target metrics.
The model updates and predicts which parameters are most likely to reach your targets in the next run.

Same protocol. Same cells. Different result.

Transcript

Same protocol. Same cells. Different result. Reproducibility is the biggest problem in biology.

ProviGen finds the cause and hardens your protocol. We connect active learning to the instruments your lab already owns, and turn process development into one closed loop.

First, we map your workflow and find where the variance comes from: donors, reagent lots, plate position, timing, noise.

Then we optimise. Every batch is chosen based on previous results, for robustness and for the targets you set: yield, purity, titre, cost, cell identity. This opens up the ability to adapt the protocols in real time.

Your scientists approve every batch. You keep the hardened protocol, a validated operating range, and the full record behind it.

We’re on a mission to make biology reproducible, one process at a time. Let’s start with yours.

Team

Strategy, operations, fundraising. Background in medicine × biomedical neuroscience, with genome editing research at Harvard Medical School and Boston Children’s Hospital. Brings a blend of business and product perspective. Studied or worked at: Harvard University, Technical University of Munich, LMU Munich.
AI, automation, product. Background in medicine × computer science × neuroengineering. A polymath who built drone swarm software, pursued stem-cell research and treats biology as a reverse-engineering discipline. His driving belief: people should be able to choose and modify their own biology, rather than have it determined at birth. Studied or worked at: Technical University of Munich, LMU Munich.
Biotech go-to-market, partnerships, scale-up. Trained molecular biologist with 8+ years in life sciences and a track record building early-stage health ventures, translating bench research into commercial propositions. De-risks commercial adoption through CDMO/pharma access and pilot-to-partnership strategy. Studied or worked at: Imperial College London, University of Wisconsin–Madison, Valneva.