Neural imaging
Custom optics combined with unsupervised learning to read the activity of neurons directly, without dyes or labels.
71% of cells
Nano Letters paperApplied AI · Natural sciences
Machine learning for research that doesn't fit off-the-shelf tools.
Thinktropy helps research groups and R&D teams bring machine learning into spectra, imaging, measurements and literature. Founded by a physicist and chemist with deep AI expertise.
- Nature Communications
- SERS + ML to identify microplastics
- 71%
- neuron activity detected without labels
- 31%
- faster literature search with AbsClust
Projects
Selected work
Microplastic identification
Low-cost SERS sensing paired with a custom model that identifies five common polymer types without sample pretreatment.
Nature Communications
Nature Communications paperAbsClust
LLM-powered discovery across the scientific literature, later applied to industrial documents.
31% faster literature workflow
absclust.comMade by Thinktropy
Free tools for researchers
doir
DOIs in, formatted references out. Turn bracketed DOIs in a Word manuscript into formatted citations and a reference list in any CSL journal style.
Free
Use doirServices
Where Thinktropy helps
“You have data and a question, and off-the-shelf tools don't fit.”
From the first question to a working method: AI strategy, custom models for spectra, images and measurements, the data pipelines around them, and workshops that help your team carry the work on.
Project consulting
“You're writing a proposal and need an AI partner.”
An AI partner for grant applications and joint projects: shaping the machine-learning work package, writing it up, and doing the work once it's funded.
Research & grant partnership
“You need an independent read on a deep-tech or AI claim.”
A scientist's assessment of whether a method, dataset or product claim holds up, for investors and R&D leads.
Technical due diligence
Founder
About
Dr. Andrii Trelin
I trained as a natural scientist and hold PhDs in physics and chemistry. Across 40+ publications, my research has paired laboratory measurements, from Raman spectra to optical imaging of living cells, with machine learning built for that specific data.
I also founded and ran a startup that brought language models to scientific literature, and have spent years on applied AI beyond the lab. Thinktropy is how I put that experience into new research and R&D projects.
- PhD Physics
- PhD Chemistry
- 40+ publications
- Startup founder
Selected publications
- ChiSCAT: Unsupervised Learning of Recurrent Cellular Micromotion Patterns from a Chaotic Speckle Pattern
Nano Letters 24, 12374–12381 (2024) · doi:10.1021/acs.nanolett.4c02425
- Pretreatment-free SERS sensing of microplastics using a self-attention-based neural network on hierarchically porous Ag foams
Nature Communications (2024) · doi:10.1038/s41467-024-48148-w
- Label-free surface-enhanced Raman spectroscopy with artificial neural network technique for recognition photoinduced DNA damage
Biosensors and Bioelectronics (2019) · doi:10.1016/j.bios.2019.111718
- Precise cancer detection via the combination of functionalized SERS surfaces and convolutional neural network with independent inputs
Sensors and Actuators B: Chemical (2020) · doi:10.1016/j.snb.2020.127660
Contact
Have data that off-the-shelf tools can't handle?
Describe the problem and the data you have. Replies within a few working days.
Read by the founder, not a sales team.