MLOps everywhere!
prokube is an integrated MLOps platform based on Kubernetes, Kubeflow, Gitlab and many more best-of-breed open-source solutions, running anywhere you need it to. On-premises, in the cloud, or in a corporate data center, prokube supports all environments you need to run your ML workloads in.
Our Main Services
With our suite of services centered around our MLOps platform, we aim to assist you in seamlessly integrating MLOps into your existing processes and tool landscapes.
Deployment &
Customization
Deploying prokube is fully automated, yet custom solutions may be needed for your specific environment. We tailor our platform to suit your application, whether on-premises, in the cloud, or hybrid.
Maintenance &
Support
Maintaining a large suite of MLOps tools, such as Kubeflow and MLflow, can be challenging at times. At prokube, we gladly take on this task for you, allowing your data scientists to focus on what they do best.
MLOps &
Data Science Consulting
We built prokube because we’ve personally experienced the challenges of developing and maintaining ML models. We are happy to leverage our experience to assist you in implementing MLOps or ML use cases.
Our Product
prokube is designed for ML use cases, based on state-of-the-art open-source technology, and is highly customizable.
Data Science Workbench and MLOps
prokube was developed to enhance the efficiency of data science teams, integrating trusted tools such as Kubeflow, Dask, Jupyter Notebooks, and VSCode. With the seamless integration of MLflow into our Kubeflow framework, prokube offers an interactive environment that streamlines machine learning workflows from end to end. It supports the effective sharing and merging of computing resources, memory, and accelerators, enhancing collaborative efforts and facilitating multi-GPU and multi-node training sessions for complex neural networks like LLMs. KServe enables scalable and reliable model serving, including for demanding models requiring high-throughput and low-latency inference. The platform’s capabilities also include continuous model re-training, automated redeployments for seamless updates, and support for custom containers, which provides the flexibility to tailor solutions to specific project requirements. These features collectively accelerate the transformation of data into actionable insights and expedite model deployment.
Data Storage and Streaming
Managing storage efficiently and effectively is a crucial aspect of any Data Science or MLOps platform. prokube offers storage and streaming solutions for a variety of use cases, such as OpenEBS and Mayastor for Kubernetes Persistent Volume Claims (PVCs) and MinIO for S3-compatible object storage, offering high performance, scalability, and data protection. Recognizing the importance of data security, Prokube offers encryption across all its storage solutions to maintain data integrity and privacy. To support streaming data, prokube integrates Kafka, which guarantees real-time data flow and processing. In addition, we include PostgreSQL and MongoDB, two powerful databases, to handle structured and unstructured data respectively.
IAM and LDAP & AD Integration
Secure enterprise login integration is critical for maintaining operational consistency and robust security. prokube addresses this by utilizing Keycloak and Dex. Keycloak provides secure, scalable access control and a range of features including Single-Sign On (SSO), and identity brokering. Dex adds an identity layer with LDAP & Active Directory support. Together, these tools enable prokube to offer a comprehensive IAM solution with robust user federation and seamless SSO.
Infrastructure as Code & GitOps
Reproducibility is key in modern development practices, extending beyond application code to include the platform and infrastructure it operates on. prokube incorporates this concept, using Ansible playbooks to ensure the entire setup, from configuring the hosts to setting up the Kubernetes cluster, is completely reproducible. In addition to Ansible, Prokube also utilizes Argo CD to implement GitOps practices, enabling continuous deployment that is both declarative and automated. This approach guarantees a consistent deployment of infrastructure and enhances scalability. Furthermore, Prokube supports both managed Kubernetes services and various on-premises Kubernetes distributions, ensuring it fits seamlessly into any infrastructure environment. This approach guarantees a consistent deployment of infrastructure and enhances scalability.
Repositories, Registries, and CI/CD
With prokube, we provide an efficient platform for managing your code, artifacts, and CI/CD pipelines. The community version of GitLab comes included, providing a centralized hub for your project. However, prokube is also designed to integrate seamlessly with any version control or CI/CD solutions currently employed in your organization.
Monitoring and Logging
Maintaining full visibility of your processes is critical. That’s why prokube includes a comprehensive monitoring and logging stack, ensuring you always have insight into your operations. Our platform comes equipped with extendable, pre-configured versions of industry-leading tools: Grafana for interactive data visualization, Prometheus for system monitoring, and Loki for centralized logging. Additionally, by leveraging KServe, prokube enhances its monitoring capabilities to include specialized features for tracking the performance and health of ML models in production, offering real-time analytics to facilitate rapid issue resolution and optimization.
About us
Born from a research venture with the Helmut Schmidt University, prokube stands today as an innovative spin-off of JUST ADD AI. Our platform is the answer to the intricate real-world challenges we’ve encountered firsthand in machine learning deployment. At prokube, we practice what we preach — our technology doesn’t just serve our clients; it’s integral to JUST ADD AI’s day-to-day operations, ensuring that every feature is battle-tested and business-ready.
Our Founders
Dr. Christian Geier
CEO & Co-Founder
With a PhD in Physics, his background includes maintaining distributed computing clusters and contributing to open-source projects, ensuring prokube is engineered for performance and scalability.
Henrik Steude
Data Scientist & Co-Founder
Henrik, our ML expert, is currently finishing his PhD in ML for cyber-physical systems. He has years of experience as a data science consultant, during which he contributed to AI system development, including work for the International Space Station. During this time, Henrik became a Kubeflow enthusiast.
Martin Creutzenberg
MLOPs Engineer & Co-Founder
Martin Creutzenberg is our Kubernetes and software engineering expert. His deep dive into MLOps tools and architectures ensures that prokube is not just cutting-edge but also user-friendly, allowing seamless integration into any production environment.