Instant Clarity: Insights for SRE Excellence
summary
Demoed application Neurologic ingests 10 GitHub repos (mix of public and private), builds up-to-date tech docs in ~2 minutes after ingestion deltas, and visualizes flows with exportable mermaid diagrams for authentication paths. Responses come from RAG pipelines and are scored by a second LLM using the RAGAS faithfulness/relevancy metrics, targeting ≥85%; observed degradations trigger delta reingestion. Long repo ingestion peaked at 53 minutes and cost about $150; 47% of users submit single-word prompts.
AI-written summary · about
chapters (9)
- 0:00Introduction and Agenda
- 0:55Overview of Neurologic Application
- 1:41Use Case: Source Code Analysis
- 4:36Visualization and Authentication Process
- 6:48Importance of Metrics and Monitoring
- 12:53Challenges and Solutions in AI Implementation
- 15:42Generating Technical Documentation
- 17:55Good Prompts vs. Bad Prompts
- 21:20Conclusion and Q&A
related talks
42mBuilding Production RAG Systems for Health Care Domains
22mThe Power of RAG at Scale: A Year of Explosive Growth and Observab...
25mSet Up an Open Source RAG in 30 Minutes
27mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
41mFrom RAGs to riches: Build an AI document interrogation app in 30 mins
40mHow [not] to evaluate your RAG
59mRAG Against the Machine
40mRolling out Enterprise AI: Tools, Insights, & Team Empowerment