Interested in contributing to GSOC 2026 😎

[calebjubal]

Hi, I’m Caleb Chandrasekar, a 3rd-year undergraduate student in Computer Science. I primarily work with Python, Java, and JavaScript/TypeScript, and I’ve recently been spending time understanding and exploring open-edge-platform.

I’m very interested in contributing to and potentially preparing a proposal around it. I wanted to ask if there are any prerequisite tasks for me to the following mentors:

• @14pankaj, @bharagha (Agentic GraphRAG)
• @14pankaj, @krish918 (User Feedback–Led Fine-Tuning of Agentic GraphRAG)
• @scottmbaker, @effndc (AI-based Edge Node Diagnostic Tool)

as the project ideas for 2026 says:

starter issues, or recommended areas you’d suggest I should work on before drafting a proposal.

Thank you for your time, and I’d really appreciate any guidance on how best to get started.

[scottmbaker]

Hello Caleb, for the AI-Based Diagnostic tool, I would start by reviewing our developer documentation here: https://docs.openedgeplatform.intel.com/edge-manage-docs/dev/developer_guide/observability/arch/index.html. This talks about our existing observability stack. The key takeaways are the tools that are used collecting logs and metrics from the edge node. Centrally we use some Grafana tools that include Mimir for storing metrics and Loki for storing logs. From that page you can drill down into https://docs.openedgeplatform.intel.com/edge-manage-docs/dev/developer_guide/observability/arch/edgenode/hosts.html and https://docs.openedgeplatform.intel.com/edge-manage-docs/dev/developer_guide/observability/arch/edgenode/clusters.html and https://docs.openedgeplatform.intel.com/edge-manage-docs/dev/developer_guide/observability/arch/edgenode/apps.html. I would think OpenTelemetry is one of the key technologies to learn about and understand. Basically, my suggestion as a prerequisite would be to understand the technologies that collect metrics and logs from an edge node, as well as the type of data that you will find on an edge node.

[bharagha]

Hi Caleb,

For Agentic GraphRAG, I would recommend starting with the following:

1. Understanding of GraphRAG. There is enough good learning materials on the web.
2. There is a semantic similarity based simple RAG implementation here. Objective of this project is to extend it with GraphRAG using sample datasets as highlighted in the project description. A prerequisite is to understand the ChatQnA implementation to be able to enhance it.

Reach out for any further questions.

[effndc]

Regarding AI-based Edge Node Diagnostic Tool since we do use Grafna, one path to investigate may be the use of mcp-grafana as a path to expose the current metrics and logs stores to an Agent. It may be necessary to add some additional context/guidance on how to diagnose various failure types that we see on edge nodes so that it knows what logs to look for.

Some of those common issues should be represented in our Troubleshooting user docs that could help as a starting point.

[biapalmeiro]

@calebjubal FYI, all contributions to Google Summer of Code will be under OpenVINO this year. Please, check the updates from there: https://github.com/openvinotoolkit/openvino/wiki/Google-Summer-Of-Code