Serokell Labs
Innovative solutions with
solid foundations
The research
division of Serokell
Our work lies at the intersection of all the important fields of modern technology
language theory (PLT)
systems
Intelligence
Our approach combines innovations from multiple fields, making it easy to create bridges between disciplines.
With our experience in machine learning, blockchain, and programming language theory, we build novel solutions to real problems plaguing the industry.
Careful research, mathematical modeling, and interdisciplinary work are at the foundation of everything that Serokell Labs stands for.
Let’s discuss your project
Fill the formAt Serokell Labs, our goal is to change industry through science
- Creating a platform for promising scientists to do original research.
- Validating state-of-the-art industry solutions within frameworks of thought already used in academia. (papers, theories, mathematical models)
- Building new solutions that take in mind the research done before in the field.
research
solutions
Path for innovators
Collaboration between Serokell developers and scientists from the leading world universities enables us to translate the vision of pioneers in optical design, computer vision, and engineering into working solutions for real business problems.
Our papers
Ethical Implications
of Brain Computer Interfaces
We provide a positive solution to the finite representation problem for representable residuated groups.
Applied research for GHC
We refactor and improve parts of GHC to move Haskell closer to practical dependent types, implementing features such as standalone kind signatures.
Research on relation algebras
We provide a positive solution to the finite representation problem for representable residuated groups.
Research on non-classical logics
We investigate topological and algebraic aspects of non-classical logics and structures that matter from a logical perspective.
Research on unidirectional coercibles
Coerce is a Haskell feature that allows for zero-cost conversion between generative type abstractions and their base type. We want to create the possibility for safe unidirectional coerce for newtypes with a guaranteed invariant.
ML Lab
Our fields of work include: applied discriminative and generative machine learning models, alternate implementations for ML algorithms with functional programming languages, coordinate-free linear algebra.
Recent projects from our portfolio:
We are investigating the use of ML to artificially increase the resolution of commodity microscopes. For the same project, we are also creating an organelle classificator that uses the U-Net network architecture.
We are exploring music generation through both symbolic representation and raw audio. Some of the tools we use: Markov chains, hidden Markov models, deep learning with RNNs and Transformers.
We are building face recognition solutions with machine learning models that use Gaussian Processes for handling high-dimensional data.
Innovation is one click away
We would be happy to discuss your project and propose solutions.
Contact us