Talmo offers customized solutions in the intersection between machine learning and software engineering. The technologies and methods used are generic and selected for the different application domains and business areas.
C++17/20 cross-platform library and application design and development with focus on current best practices and keeping up with the changes to the C++ standard and related build, testing and packaging technologies like CMake, Catch2, Conan, clang-tidy etc.
Development of custom models in both Tensorflow and Pytorch for solving identification, regression, classification, detection and segmentation tasks in projects. Special focus on the architecture trade-offs considering both training feasibility and inference
GUI development for both Windows devices using C#/.NET and cross-platform applications based on the Qt application framework for Windows, Linux and Mac.
Designing and development of interfacing approaches between native C++ libraries and managed client applications written in languages like Java, C# and Python considering both performance and ease of use from the client application.
Advising about best practices for testing software containing computer vision and machine learning technology in general, i.e. how to deal with large sets of reference data with a requirement for fuzzy comparison due to the stochastic nature of some methods.
Development of custom machine vision applications based on commercial packages like MVTec Halcon, Cognex Vision Pro and Stemmer Common Vision Blox.
Android and iOS development with special focus on machine learning and computer vision apps. Interfacing of native C++ libraries like OpenCV, Ceres and Eigen for usage through the respective app platform technologies.