Building Your Agent Indexing System
with Doug Turnbull
Register here. It’s a lightning lesson in RAG: Retrieval Augmented Gathering, the free series Doug Turnbull runs on Maven, at 1pm EDT on Friday September 25th. Free to attend, recorded afterward, and it may contain a promotional CTA for hevlayer.com.
What you’ll learn
How observability standardized, twice. Metrics consolidated on Prometheus. Tracing consolidated on OpenTelemetry. Both times the fix was an agreed format.
The terms of art: trace, span, metric. The vocabulary you need to direct an agent to build this, and the naming conventions that make harnesses comparable.
Where the span model stops. Spans assume short, thin, and disposable. An agent session is hours of prose you want to keep and search for years.
Why you want warehouse storage. Keep the raw capture beside a versioned Parquet view, so a schema change becomes a rebuild instead of lost history.
Hybrid search over a real trace archive. Live CLI and UI over six months of my own agent sessions, and what the index actually costs per month.
Why this talk matters
A coding agent reads more of your repository in an afternoon than a new hire reads in a month. It writes down every file it opened and every approach it abandoned, and then the session closes and all of it goes in the bin. What survives is usually run counts and token spend, which is enough to bill against and too thin to engineer against. This talk builds the archive instead.