Blindspot Solutions

In many situation the companies need to optimize the utilization of their employees and assets, being it vehicles for the delivery of goods, auditors and inspectors for audits or processing time of the processors at the computation units. Leveraging our knowledge in mathematical optimization and machine learning, we can provide customized algorithms providing the most efficient solutions, scaling to thousands of units and able to find cost-efficient schedules in minutes.

Typically, the designed algorithm can be expressed as a mathematical program and we use either commercial solvers, such as IBM CPLEX to find a solution or we use one of the free open-source solutions, such as OptaPlanner, GLPK and others. In some cases, we deliver a tailored algorithm, designed for the specific problem, thus vastly outperforming generic solvers (both free and commercial).

One of the leading global notified bodies requires to schedule its inspectors in various regions around the world. We designed a customized algorithm taking into account different constraints of different business lines. The algorithm is able to schedule activities of hundreds of inspectors visiting thousands of places; all within minutes. The solution helped to dramatically save travel and accommodation costs as well as to increase the utilization of the inspectors.

We are following agile development practices and we adhere to the highest coding standards with a strict code review and test coverage. We deliver on time and we support frequent iterations with the client or software users. We help the clients to define their needs by suggesting requirements and often jointly preparing software specifications. We do integration tests on site and we provide support based on client needs.

Our software typically contains a module solving either a machine learning or an optimization problem described above. Typically, we are using Java / Scala stack. In case of prototyping, we heavily utilize Python and its numerous libraries. For frontend, we alternate between AngularJS and React. We have built a few Java desktop application with JavaFX frontend and for a Microsoft-locked customer, we developed a system in C# .NET. Deployment done both locally and in the cloud (AWS, IBM Bluemix).

For Crocodille s.r.o., we have developed a Java application which made sense in a huge amount of logistic data stored in a database organically growing over many years. The users' requirement was a desktop application with rapid response, platform independence and a beautiful user interface. Java with JavaFX frontend allowed us to design custom controls for drill-down-based logistic process exploration and AI-aided optimization module was integrated to propose cost-efficient supply routes adhering to tight delivery constraints. Manual reconfiguration of the routes was made possible by advanced hint system providing the user with an instant response after each modification.