Fern IT

A BärnHäckt 2025 project.

Mit KI zu GreenOps - Digitalisierung wird ökologisch nachhaltiger

(With AI to GreenOps - Digitization becomes environmentally sustainable)

CONTEXT AND BACKGROUND

The digital transformation is leading to a steadily increasing energy consumption in IT. At the same time, developers and operations teams often lack tools and support to understand and actively reduce the ecological footprint of their software. While the specification software carbon intensity (SCI), for example, is an important benchmark, there is a lack of integrations in everyday software development and operations. The challenge is to bring concepts, data and measures closer together for more Green IT. We not only need dashboards that visualize CO2 emissions, but also intelligent tools that support making ecological decisions and implementing measures.

DESCRIPTION OF PROBLEMS

Task / Objectives:

Your mission is to design an intelligent assistant that helps developer and operator teams to sustainably optimize the Carbon Intensity (SCI) software throughout the entire life cycle of a software.

Frame / specifications:

• Technology stack : Choosing the technology and AI stack is free as long as the solution becomes largely open and interoperable. Frameworks for chatbots and programming languages (Java, Ruby, Python, Go, JavaScript/TypeScript) are welcome.
• Open Source & LLMs : The solution should be as open as possible, prefer to use open source LLMs and be published as open source.
• SCI Calculation : Use the information of the Green Software Foundation as a basis (ISO standard).
• Output : A clever concept, a cool presentation and a tangible solution prototype – that's not just what you score with BärnHäckt.

Additional ideas for implementation:

The following aspects are taken up with intelligent, AI-powered assistants:

• Repository analysis & recommendations: The wizard is able to analyze Git repositories and, based on code, used languages, build processes and dependencies, to generate concrete recommendations to reduce energy consumption and resource usage. Some examples:
  • Suggestions for code optimization (more efficient algorithms, reduction of I/O operations).
  • Recommendations for leaner container images.
  • Notes on optimizing CI/CD pipelines (avoidance of unnecessary builds, optimization of test runs).
  • Proposals for the use of more energy-efficient libraries or frameworks.
  • Architecture tips for more sustainability (e.g. Event-Driven Architectures, Serverless).
  • Optimization of batch running based on the availability of green energy.
• SCI Monitoring & Recommendations (Kubernetes/OpenShift): The wizard can connect to a Kubernetes environment (ideally OpenShift), captures real-time metrics (e.g. CPU usage, memory-use, network traffic) and calculates/visualizes the Software Carbon Intensity (SCI) of ongoing applications (in the case of missing information, assumptions are used or they can be entered interactively (e.g. embodied carbon and CO2e per KWh). Based on this live data, the wizard makes recommendations. Some examples:
  • Recommendations for automatic scaling of pods based on actual needs.
  • Identification of workloads that consume unnecessarily many resources and suggestions for optimization.
  • Notes on shifting workloads in times with greater green energy availability.
  • Contextual recommendations for adjusting resource limits and quotas.
• Interactive communication & learning ability : The AI assistant communicates interactively in a chat, asks questions, makes concrete and proactive suggestions and discusses with the user (developer and operator). Ideally, the assistant can learn from these interactions and improve recommendations about the time.

Challenge Sponsor: Puzzle ITC

Puzzle ITC is an established software and technology company with a 25-year history. More than 140 committed people are implementing sophisticated IT projects at its Bern, Zurich, Basel and Tübingen (DE) sites. The focus of our business activities is on open standards and economical, resilient IT. Companies and organizations receive competent support – holistic or as needed from the initial idea to the operation of a digital solution. Business analysis, project management, the own user experience studio, software development as well as system, container and platform engineers work hand in hand. We develop and implement agile, use modern open source technologies and accompany them into the new, rapidly evolving world of AI. Cloud Native applications and modern, future-proof IT solutions are created tailored to the production.