Fluorescence data correction platform for a biotech startup

Fluorescence assays are a workhorse of modern research. Pharma and biotech R&D teams, contract research labs, and quality control departments run them on microplate readers every day.

They share a stubborn problem. At higher concentrations, the sample itself absorbs part of the light the instrument relies on. This inner filter effect bends results that should be linear, and the usual fixes are slow and costly: dilute the samples, rerun the plate, or throw away the top of the range.

A reliable correction is valuable science. But it only changes how labs work when researchers can use it on their own data, without a physicist in the room.

TINFE is a spin-out from the University of Zagreb Faculty of Pharmacy and Biochemistry, founded by experts in fluorescence spectroscopy and computational chemistry. Their physics-based method removes inner filter effects directly from microplate reader data, with no extra sample handling, no hardware changes, and no reruns. The company is backed by Vesna VC, supported through European and national innovation programs, and took an award in ZICER's Tech Transfer program at Zagreb Connect.

Their correction engine was proven. What it needed was a platform that labs in pharma, biotech, academia, and industrial QC could sign in to and use.

That's harder than it looks. Every instrument manufacturer exports data differently, calibration has to be right before anything else can be, and researchers won't adopt a tool whose results they can't inspect. TINFE also needed to keep full ownership of the science, and keep improving it, without the platform getting in their way.

We started with UX design, turning a strict scientific process into a path a first-time user can follow.

Users begin with calibration. The platform recognizes calibration values in the uploaded files when they're present, lets the user define them when they aren't, and shows the result, which can be saved and reused later. Then users upload their measurements, the platform runs the correction, and the results appear on screen, ready to download.

The most important decision sits under the surface. TINFE's own correction engine runs as an external resource, cleanly separated from the application around it.

The platform handles everything else: security, user management, file handling, storage in Amazon S3, orchestration, and presentation. It converts each upload into the format the engine expects and turns what comes back into something a researcher can read. TINFE can keep refining the science without touching the web application, and the product is delivered turnkey.

The first version read data from two instrument manufacturers. Since then we've added parser modules for different measurement platforms, each translating a vendor's file format into one shared structure.

That matters commercially as much as technically. Every instrument the platform can read opens a new group of labs to sell to, and adding one now means writing a parser, not reworking the product.

Researchers need to see what a correction did before they'll trust it. So alongside the tables and explanations, results come with interactive 3D graphs built with Plotly that users can rotate and inspect. Instead of a number to believe, they get a picture of their own data to explore.

The platform was also built with growth in mind. It logs user actions, so TINFE can see how researchers really use the tool, and its architecture is ready for company accounts when seats need to be managed per organization.

TINFE's platform is live in production and open to demos for labs that want to see it working on real plate reader data. Researchers go from a raw plate reader file to corrected, explorable results in a single session.

Our work comes in phases, and it's still going. Each time the product needs to grow, we pick it up with full context of what's already there. For a science-led company, that's the point: the founders stay focused on the physics, and the platform keeps pace with them.