Projects

With a proven track record and a forward-looking mindset, KubeSpectra has successfully delivered impactful projects across various industries. Our past projects reflect our expertise in platform engineering, private and public cloud solutions, software engineering, and AI. From architecting scalable infrastructure to developing intelligent applications, we have consistently exceeded expectations. In our current projects, we continue to innovate, leveraging the latest technologies to drive digital transformation and deliver tangible business outcomes. Partner with KubeSpectra to unlock the potential of your projects and propel your organization towards success.

Education

First lecture on cloud computing in Germany

We created the first lecture on Cloud Native Comuting in Germany. Read more

Fin tech

OpenShift Platform

The goal was to install Cloud Pak for Data in Openshift OKD. For this we used Linux VMs from Azure.

Public service

Rancher Kubernetes Platform

Situation:
Migration of internal applications to a new intranet platform, with a focus on open source software and a cloud native architecture.

Project approach:
The objective was to implement an overall architecture for platform, local development environment, network and all interfaces (users, databases, storage, etc.). The focus was on the introduction of:

• Kubernetes and a cluster management system
• communication encryption / access control via service mesh
• introduction of new open source tools
• DevOps concepts
• Operators and other automation solutions

In addition, the developers were advised on migration with regard to cloudnative application architectures. For this purpose, general guidelines and architecture decisions for new technologies were created and blueprints for microservice architectures, claim check patterns and monitoring of performance and availability of services were developed.

Cloud native AI platform

Based on KubeFlow and with a GPU connection, an AI platform was developed in a Kubernetes cluster. Models are to be developed in this platform which, according to the security concept, must not leave the Kubernetes platform (development environment and virtual computer in the platform).

Research & Development

Docker without desktop

Docker Desktop becomes commercial, we show you a free alternative - Docker in Windows Subsystem for Linux (WSL). Docker's free products are used by millions of developers to build, publish and run applications - in data centers, the public cloud or with Docker Desktop on the local PC. 55% of developers use Docker every day at work. Read more

D/A trading

Data-driven trading recommendations in D/A power trading. This is based on market data that is analyzed using state-of-the-art AI algorithms. From this, forecasts are calculated and made available via online dashboard or messaging system. With these recommendations your trading performance increases and at the same time it reduces the time for market observations.

Peak load forecasting in the energy market

Load peaks occur in many networked systems, such as data centers, logistics or power grids. In the case of internal power grids, load peaks result in disproportionately high costs, since either excess capacities have to be maintained or services have to be purchased externally. Using AI-based methods, energy load peaks (which cost a lot of money) are automatically predicted and intercepted via adapted production control. This is done by machine learning, which is used to create transparency about the max. daily load withdrawals. The fine control (management) of the decisive energy consumers can thus be optimized.

Conclusion: Especially medium-sized industrial companies can significantly reduce their energy costs with this lightweight AI solution - quickly, easily and safely.

OPServer

OPServer is a maintenance tool for platform operators and is used for continuous monitoring of PaaS systems. For this purpose, a series of small testers were implemented, which, for example, check whether external access from load balancers is possible. The tool then collects the test results including logging and monitoring statistics and shows in a UI if everything works. The Git repo with the source code can be accessed here Github OPServer.

Skill database

A web application where users can enter their technical skills. This information is sent from the frontend to the backend and from there to the graph database "Neo4j". Afterwards, the aggregated data/user information can be viewed in the Shopify format. The goal of this application is to be able to assign suitable employees to projects/project tenders as quickly as possible and to have contact persons for individual technologies.

Test data generation

Before new applications go into production, data-driven tests are essential to ensure the quality of software and applications. For these tesings we have built a test data generator. Read more

Video Analysis

Development of an AI system that uses a neural network (ssd-mobilenet-v2) to analyze the video feed of a camera and extract characteristics such as number of people, direction of movement, clothing, etc. These characteristics are further processed by a subsequent cloud system. These characteristics are further processed, aggregated and visually displayed by a subsequent cloud system.

Web Crawler

WebCrawlers are a simple, effective and inexpensive way to search websites for specific information and make it available in compressed form. The programs are thus ideally suited to perform repetitive tasks. Read more

Sports

Sports analytics with FC Bayern München Basketball

Unlike other sports, basketball has been using statistical analysis for decades to improve game, player and team coordination. We have further developed these approaches together with a top professional club and by using advanced analytics. Read more

Software distributer

AI assisted planning

Development of an (AI-based) "intelligent assistance function" to complement existing planning software for industrial applications. With the currently used planning software, the user has to set a fixed set of parameters and can then, through manual, iterative changes, achieve a result. This "trial & error" approach is to be supplemented/replaced by an automatic optimization tool.

Top domain provider

Concept for traffic data, analysis, security and alerting

We created a concept with requirements how Kubernetes can be used in combination with Clickhouse and appropriate evaluation tools to realize a performant platform for a large amount of data.