Articles for people who
ship the stack
Rewrites on React, Node.js, TypeScript and AI. A translated body is shown when it exists. If it does not, you see the English text and a note that says so.
Tagged: Debugging
When the Dashboard Is Green and the Agent Is Wrong: Tracing State
The dashboard is green and the agent is wrong: failures hide in state passed between healthy steps. What tracing, routing records and evals must capture.
12 min readDiagnosing RAG Failures: A Stage Map and a Claim-Level Groundedness Judge
Localize where a RAG pipeline broke, then build an LLM groundedness judge that checks claims one by one and is validated against human raters with kappa.
9 min readHow a Docs-Fixing AI Agent Fooled Itself: False Positives in a Docker Sandbox
A LangGraph docs-fixing agent shipped a no-op fix, then broke on src-layout packages. The Docker sandbox bug behind it and the real fix it found in click.
7 min readDebugging Quiet RAG Failures: Retrieval, Chunking, Filters and Evals
RAG answers can sound right and still be wrong. A Python-minded playbook for tracing retrieval, chunking, metadata filters, evals, hybrid search and grounding.
8 min readSeven Silent Ways Context Pollution Breaks LangGraph Agents
Stale state, giant tool dumps, duplicate RAG chunks, lossy summaries and bloated toolsets: seven failures that raise no errors, with fixes and how to detect each.
9 min readWhen an AI Agent Fails, Who Do You Blame? Tracing the Failure Stack
Why 'the model hallucinated' falls short: a walk down the agent failure stack from intent and retrieval to tools, retries and traces, and what to observe.
8 min readSeven Ways a RAG Answer Goes Wrong and How to Find the Failing Stage
A practical way to diagnose wrong RAG answers: seven failure points grouped into three stages, four fields to log, and one test that names the culprit.
7 min readDebugging Serverless Apache Spark with Gemini, MCP and Agent Skills
Debugging serverless Apache Spark with Gemini: from pasting tracebacks into chat to an agent that reads batch logs via an MCP server and fixes PySpark.
6 min readFinding Why a Weaviate Query Is Slow With Per-Stage Profiling
Weaviate query profiling returns a stage-by-stage timing breakdown with each response, so slow filters, disk reads, graph search or rescoring can be told apart quickly.
5 min read
About these articles
Request a 24h estimate
Need the same stack in a production operator layer? Send the brief — estimate within 24 hours.