Making tool-calling agents reliable on AWS Bedrock
Rate limits, timeout cascades and tool schema drift took down our Claude agents more than any bad prompt did. How we fixed each one and reached 99.9% uptime.
Writing
Notes from building agents and retrieval systems
What broke in production, what fixed it, and the patterns I keep coming back to. Written by Sushank Ghimire, mostly late at night, always after the incident is closed.
RAG on messy HTML: preprocessing, structure-aware chunking and hybrid retrieval
Fixed-size chunking fails on real documents. What moved retrieval quality for us: boilerplate stripping, heading-aware chunks, hybrid search and a re-ranker.