BITMARCK
New data science platform accelerates the AI transformation of health insurance providers
At a glance
Client: BITMARCK, a leading digitalisation partner for statutory health insurance – more than 80 per cent of German statutory health insurance providers are clients of the company.
Sector: Healthcare
Challenge: Project-specific AI development environments resulted in high set-up and operational costs, limited reusability, and significant regulatory hurdles for social data
Solution: Development of a centralised, secure data science platform for the standardised and easily scalable implementation of AI projects within the BITMARCK ecosystem
Results: Significantly shorter project lead times, simplified operations and a sustainable foundation for an AI marketplace
Challenge
Standardising individual development environments for the rapid and secure introduction of AI in health insurance companies
German health insurance providers are under enormous pressure to improve efficiency. Artificial intelligence is seen as key to reducing costs and improving customer service. However, as the healthcare sector deals with highly sensitive personal data, AI development must meet specific requirements regarding data security and governance.
As a leading digitalisation partner for statutory health insurance providers, BITMARCK ensures that digital processes run smoothly in the background and has, until now, implemented AI projects in customised set-ups within the secure on-premises infrastructure at the BITMARCK data centre. Whilst this ensured the highest security standards were maintained, the effort involved did not scale with the increasing demands – not only when setting up the development environment at the start of a project, but also during the operation of the applications.
Given the rising demand for AI applications, such a setup was no longer viable for the future. BITMARCK therefore turned to us to adapt its development infrastructure for secure, efficient and scalable AI development.
Solution Approach
Centralised data science platform for all AI projects within the secure BITMARCK infrastructure
Using agile methodology, we worked with BITMARCK to develop a central data science platform on which both internal AI projects and client projects can be efficiently implemented and operated end-to-end – whilst maintaining full data sovereignty.
-
The platform itself is hosted in BITMARCK’s data centre. We utilise an air-gapped environment, meaning access is only possible via VPN or a terminal server with a virtual desktop from within BITMARCK’s secure network. Based on selected whitelists, the infrastructure restricts access in accordance with the Zero Trust model to what is strictly necessary. Network shares to standard resources are pre-configured. Artifacts from public repositories such as PyPi or HuggingFace are mirrored and checked for vulnerabilities and integrity in order to minimise supply chain risk. Furthermore, the various client projects are strictly isolated from one another within the multi-tenant environment.
Results and benefits
A future-proof, secure setup for the development and operation of AI applications
The data science platform is currently in the early production phase: BITMARCK has already successfully implemented its first projects on it with a major statutory health insurance provider. In future, all client projects are to be implemented using the new setup. We will continue to develop the platform in collaboration with BITMARCK based on user feedback, but the benefits are already evident in day-to-day operations:
- Efficient: New AI projects can be launched technically within a day. There is no need to set up individual development environments. Reusable modules reduce the time needed for staff to get up to speed and improve development quality.
- Reliable: Platform operations are standardised and largely automated.
- Transparent: A documented tech stack provides clarity for development, operations and security – and improves compliance.
- Secure: Client isolation, controlled access and on-premises hosting meet the high requirements of the healthcare sector and significantly reduce shadow IT.
- Modular & Scalable: AI services can be expanded and operated in a controlled manner without fundamentally adapting the architecture or introducing new operating models.
- Future-proof: The platform forms the basis for an AI marketplace through which tried-and-tested AI solutions are offered to health insurance providers.
“The new data science platform is a real milestone for our AI development. We can now implement new use cases for customers much more quickly – and have created an infrastructure that will enable us to scale our capabilities in line with the evolution of AI.”
With its new data science platform, BITMARCK has established a future-proof framework for the development and operation of AI applications. Instead of individual, isolated solutions, the company now relies on a central technological foundation that accelerates innovation, reduces complexity and meets the specific requirements of the healthcare sector. This enables BITMARCK to continue developing AI solutions in a structured manner over the long term and strengthens its role as an enabler of digital transformation in the healthcare sector.
“Our collaboration with Comma Soft was consistently solution-oriented and, above all, conducted on an equal footing. This enabled us to find an efficient setup that fits in perfectly with the way we work and supports our quality standards.”
Is your infrastructure ready for the AI era?
If you’d like to organise your AI development more efficiently and achieve greater scalability, please feel free to contact us at any time.
