The need of Security On-Demand was to increase its ability to detect suspicious network events based on network traffic data. The main goal of the project was to support Security Operations Center employees in ongoing monitoring of client security. The solution we have proposed was crowdsourcing algorithm. Providing the input remained on the customer’s end. In our activities, we used the capabilities of the original KnowledgePit.ml competition platform, where our team has been successfully hosting similar endeavours for several years. QED Software was responsible for the formal definition of the problem, data preprocessing, preparation of the benchmark and competition evaluator, promotion of the competition, as well as for conducting the competition and running a conference session to summarize its results.

A total of 250 teams presented their solutions. Apart from cash prizes, the winners were granted participation in the IEEE Big Data conference in Los Angeles and the opportunity to publish a conference publication. The best of these solutions were used to create an optimized algorithm, which is the basis for implementation and further development by the client’s team.

The aim of the project was to respond to the market’s need in the form of improving the targeting quality of advertising in the largest Digital Out-Of-Home (DOOH) network in Poland.

The solution is a system that is capable of matching content to its recipients of nearby media in real time. This process is based on the analysis and prediction of demographic data. The challenge that has been overcome was to ensure the autonomy of monitoring units, while providing high quality of prediction and matching of content to its recipients.

Bloober Team is one of the most popular gaming studios that create psychological horrors. The company has noticed that the market lacks titles that use the measurement of player emotions and decided to develop its industry by creating a tool to enable such measurements and use them in the course of the game. In response to this challenge, we helped the client create a unique system to detect the emotions and behavior of players that does not require additional equipment. It allows game developers to introduce gameplay elements that automatically adapt to the affection of players.

Information Builders (IBI) is one of the most prominent players on the business intelligence market. To stand out from the competition, the company wanted to expand its R&D team with data science competencies.

The business decision was to locate the core of this team in Poland, within the framework of QED Software. This allowed IBI to promptly expand its offer with services expoiting the latest achievements of applied mathematics – both as part of predictive services, approximation inference, and the creation of advanced statistical models.

Sierra Gorda SCM (KGHM), a Chilean open-pit copper and molybdenum fossil mine, has reported the need for a mathematical model showing the extent to which air pollution in the mine’s surroundings is a result of mining activities.

Mathematical modeling and data analysis of the environmental impact of the mine’s operation have allowed to assess the real impact of its extraction for air pollution in the area. This type of analysis is used in the conscious management of the level of extraction, for example to increase it without adversely affecting environmental pollution in residential areas.

Zoined is the creator of one of the most interesting business intelligence solutions targeted at retail networks. Their platform allows network owners for an instant insight into a synthetic approach to sales data, as well as convenient analysis reaching the level of individual checkout points. The company wished to enrich the panel and provide its clients with a predictive module to predict sales and demand for their products.

The predictive engine we created has been test-implemented in a nationwide Swedish bakery network and a clothing stores network. The success of these implementations convinced the client to include it in the services offered along with the Zoined Analytics platform.

Contemporary companies in the telecommunications industry are struggling with the problem of rising costs of customer service – especially in the face of the psychological barrier and reluctance of customers to be served by machines. This results in increasing pressure on the cost reduction of call centers’ services, without compromising their quality.

An interesting solution that can be used in every industry in which there is an element of customer telephone service is the product developed by SilverWhisper – an innovative product listening to an ongoing conversation, and based on its context prompting customer service employees and presenting them with materials that they currently need, or will shortly need in the course of their conversation. This reduces customer service time and increases its efficiency – positively affecting customer response and evaluation.

The role of QED Software in the project is to provide expertise on the creation of effective and interference-resistant inference algorithms – the artificial intelligence that is the core of the product.

The need of Destructive Creation, one of Poland’s most thriving computer game developers, was to create a strong and realistically behaving artificial intelligence for multiplayer.

The solution was to implement the prototype version of the Grail engine in Ancestors Legacy – a solution that enables creation of advanced and easily modifiable behaviors in multi-agent simulation environments (including video games). In addition to the high quality of the product, confirmed by positive reviews and good feedback from players, an additional value we provided for Destructive Creations was relieving their team by fully taking over the AI area.

An additional challenge was the short timeline to implement the project. The need arose at the game usability testing process, taking place during the beta testing phase.

TrichoLAB is a unique company – it is one of the few companies in the world dealing in hair and hair follicles research, in a way that combines process excellence with extensive medical and research facilities.

We had the opportunity to support the TrichoLab development team in creating machine learning mechanisms responsible for digitizing trichoscopic images and automatically creating hair and hair follicles metrics. The resulting solution was included in the company’s basic business process, enabling the reduction of patient examination costs on the one hand, and the improvement in the quality of the results of the examination on the other.

An additional element in supporting TrichoLAB was the organization of a data science competition, in which participants solved a specific problem regarding trichoscopic images. For this purpose we used the original Knowledge Pit competition platform.