HIGH-THROUGHPUT IMAGE RECOGNITION ENGINE

The client operates in a field with a strong and ongoing focus on scientific software tools for chemical analysis. Based on this, K-Softex prepared a focused service offer aligned with the company’s area of work. During the initial call, the client described the current challenges and expectations regarding a new system for image classification and recognition.

Server infrastructure for scientific image processing

PROJECT DESCRIPTION

The primary goal was to develop a server-side solution for classifying and recognizing images of chemical formulas and organic reactions. The system had to process large datasets without a web interface and support distributed execution across thousands of servers.

Core languages and runtime

Java / Python / C++

Machine learning and computer vision

TensorFlow / PyTorch / OpenCV

Infrastructure and orchestration

Jetty / Javalin / Dirk

CLIENT EXPECTATIONS

The client had high standards for both communication and execution:

Daily syncs to track progress and confirm mutual understanding
Fast development cycles
Experienced engineers capable of delivering results
Clear, relevant suggestions about technology choices, including the reasoning behind any proposed changes.

Our team at K-Softex suggested integrating proven open-source libraries and research-based models. This reduced development time and increased reliability for production use.

The core stack consisted of Java and Python. The project also involved specific hosting restrictions. This required us to build custom solutions and maintain full control over the environment.

One of our key suggestions during planning was to move from Java 8 to newer versions. Java 21 was proposed as the main development version, while some modules used Java 17 to match client infrastructure requirements. C++ was also used for parts of the processing engine.

High-performance server environment for chemical image recognition

THE RESULT

Development is ongoing, with full alignment between the engineering team and client expectations. All significant decisions have been made with open communication and technical clarity, and the project remains on schedule.

The client is pleased with the results delivered by K-Softex and is preparing to involve the team in an upcoming development project.