The Debugger: RAG Based Agentic Systems Debugger
- AI-driven troubleshooting agent that understands logs, analyzes system behavior, and even hypothesizes root causes like an engineer would.
- It can be connected to real robots via raspberry pi and mqtt protocol along with the protocol the robot is capable of communicating over.
- It uses RAG to use the documents to provide the context to the problem being asked to the debugger.
- A vector database is used to store the indexed document embeddings.
- It can also pull logs from the robot for a given query in real time and process the information to provide better resolution.
- It uses Agentic RAG along with LLM to route the query appropriately, use the tools as needed and provide a resolution response.
