Hi everyone,
My name is Aditya. I have been diving into the jenkinsci/resources-ai-chatbot-plugin codebase and have put together a draft proposal for the “Continue AI-Powered Chatbot” project.
I haven’t been as active in the community channels this month as I would like because I am currently in the middle of my 3rd-year university exams (I actually have a paper tomorrow!). However, I wanted to ensure I shared my draft with you all to get your valuable feedback.
Rather than just expanding the RAG pipeline, my proposal focuses heavily on the local hardware constraints of Mistral-7B. I have designed a deterministic Intent Router to bypass the LLM for simple queries, a sliding-window log extractor to prevent OOM crashes, and a React-based Human-in-the-Loop (HITL) UI to safely gate mutating actions.
Draft Link (Comment Access): GSOC2026_Tiwade_Resource_Chat_bot - Google Docs
I would be incredibly grateful for any high-level feedback from the mentors or community. I am specifically looking for thoughts on whether the “Auth-Agnostic Proxy Layer” I proposed for Phase 1 aligns with the team’s long-term security goals.