FAQ
Questions, answered.
Straight answers about working with us, your data and IP, scope, design of experiments, cost and timelines. Something missing? Ask it on a call.
Working together
Do you only work with microbiology labs?
No. Microbiology is where our founder comes from, and it shows in how we work with any lab: we know what a run, a plate map and a failed control look like.
We build for Life Science teams across molecular and cell biology, genomics, diagnostics, bioprocess and assay development.
We don't have a software team. Can we still work with you?
Yes, and that is the usual situation. We set up what the tool needs to run, train the people who will use it, and hand over documentation written for scientists, not developers.
From your side we need one person who knows the workflow and can answer questions.
We already have an ELN or LIMS. Will you replace it?
Usually not. Replacing a system your team already knows is rarely the best use of your budget.
We connect to it through its API, or its import tools where there is no API, and fill the gaps around it: the instrument that doesn't talk to it, the report nobody wants to assemble, the step that still lives in Excel.
Can you work with the scripts our scientists already wrote?
Yes, and it is often the fastest place to start. We turn working R or Python scripts into tested, documented tools that anyone on the team can run, and we keep the science the scripts already got right.
What happens on the first call?
Thirty minutes about one workflow: where the data comes from, where it needs to go, and who loses time in between. No slides.
You leave with a clear next step, whether or not we end up working together. If it's a fit, a written proposal with scope, timeline and a fixed price follows.
Do you work with academic labs and core facilities?
Yes. Core facilities and research groups face the same data problems as companies, often with less time to solve them. We scope the work to fit grant budgets and timelines.
Data and IP
Who owns the code?
You do. Once each stage is paid, the source code, the documentation and the IP created for you in it are yours, and that is written into the agreement.
Open-source components keep their own licenses. We don't resell a tool built for one client to another.
Where does our data live?
In your environment wherever possible: your cloud account, your servers, your ELN.
We build and test with sample or anonymized data when we can, and we only touch production data with your permission and for as long as the work requires.
Will you sign our NDA?
Yes. We're happy to sign your NDA before you share anything confidential.
The first call doesn't need one: we only need to hear how the workflow runs, not what you are discovering.
Do you build GxP-validated or 21 CFR Part 11-compliant software?
Our focus is research and development work outside GxP.
If a tool may later need to enter a regulated process, tell us at the start: we'll design it with that in mind (audit trails, access control, electronic signatures, change history) and be clear about what formal validation would involve.
Scope and technology
What kind of tools do you build?
Software that sits between your instruments, your data and your people: data pipelines, analysis apps, bioinformatics workflows, sample trackers, dashboards and AI tools for documents.
If it involves lab data and a step someone repeats by hand, it is probably in scope.
Which technologies do you use?
Whatever fits your environment and your team. Typically Python and R for analysis, Nextflow or Snakemake for pipelines, containers for reproducibility, and web interfaces for the tools scientists use every day.
We deploy to your cloud (AWS, Google Cloud or Azure) or on your own servers.
What happens after launch?
You get the code, the documentation and a short training session for the people who use the tool.
After that you run it yourselves, with us on hand. A support plan covers fixes and updates as your workflow changes, and new features get their own fixed price before we start.
Bioinformatics
Do you sequence our samples?
No. We build the analysis: the pipelines, tools and reports for data from your sequencing core, your provider or your own instrument.
If you are still choosing a provider, we can tell you what to ask for so the data fits the analysis you need.
Can you work with a pipeline we already have?
Yes. Often the fastest route is to harden what already works: add tests, containers and documentation, then make it scale. The code stays yours.
Which protein AI models can we use commercially?
Most of them. Boltz-2 (MIT license), Chai-1 (Apache 2.0), AlphaFold 2, ESMFold, RFdiffusion, ProteinMPNN and BoltzGen release their code and model weights for commercial use, and so do the open ESM3 and ESM C models, which moved to the MIT license in May 2026. The largest ESM3 models are available only through an API, under research terms.
AlphaFold 3 is the main exception: its weights may be used only by non-commercial organizations for non-commercial work, and companies use it through Google Cloud. Dependencies count too: BindCraft is MIT-licensed but needs PyRosetta, which requires a paid license for commercial work. We check every model and its dependencies before they go into your pipeline.
Design of Experiments
What is design of experiments (DoE)?
A way to plan experiments so that several factors change at once, in a pattern chosen so that the effect of each factor, and of pairs of factors, can be separated afterwards. It finds interactions that testing one factor at a time misses, and estimates each effect more precisely from the same number of runs.
In the lab it is used for media and fermentation conditions, assay development, formulation, transfection and the robustness of analytical methods.
Do we need JMP or another statistics license?
No. We design and analyze with established open-source statistics in R or Python, and you get the scripts, so rerunning an analysis needs no license.
If your team already uses JMP, keep it: we can work from its exported designs and data tables.
How many runs does a DoE need?
It depends on the number of factors and on what you need to learn. Screening 5 factors, including every two-factor interaction, takes 16 runs plus a few center points. Optimizing the 2 or 3 factors that matter then takes another 11 to 20.
We size the design to your plates, your budget and the variation in your assay.
Can a DoE support a regulated process or method?
Our focus is research and development outside GxP. Quality by Design (ICH Q8 and Q11, and ICH Q14 for analytical procedures) builds on multivariate experiments like these, so if the work may move toward a filing, tell us at the start.
We'll document every design choice and be clear about what formal validation would involve.
Cost and timelines
What does a project cost?
It depends on scope, so we price each stage as a fixed amount, agreed in writing before work starts. The first call and the written proposal are free.
Most teams start with one focused tool and grow from there once it has proven itself.
How long does a project take?
It depends on the workflow: a focused tool, such as a pipeline from one instrument into your ELN, is much quicker than a larger platform.
Larger platforms are built in stages: you get a working version early, then improvements, and you approve each stage before the next one starts.
Next step
Which step in your lab still runs on copy-paste?
Tell us in a 30-minute call. You'll leave with a clear next step, whether or not we end up working together.