▸ Topic · RAG
RAG that survives past the demo.
Demo RAG looks fine on three happy PDFs. Production RAG fails on chunking, retrieval, cost, and silent regressions. This hub collects the RAG writing and engagement paths I use with product teams.
What product teams get wrong about RAG
Most RAG failures are not "the model is dumb." They are bad chunks, weak retrieval, missing keyword fallback, and no eval gate before release.
You do not need every vector vendor on the market. Many SaaS teams can keep documents next to the app with PostgreSQL and pgvector, Laravel jobs for ingest, and hybrid search when pure vectors miss obvious keywords.
Measure before you scale. A small golden set and cost guardrails beat another prompt rewrite. Start with the posts below, then the production RAG problem page or the AI sprint when you are ready to ship.
Curated RAG reading
-
▸ Post
Why your RAG is failing
Common production failure modes and fixes that move quality.
-
▸ Post
7 RAG mistakes in production
The mistakes that keep showing up after the demo stage.
-
▸ Post
Picking the right RAG stack
How to choose storage, retrieval, and orchestration without hype.
-
▸ Post
RAG architectures compared
Traditional, agentic, and corrective patterns in plain terms.
-
▸ Post
RAG vs fine-tuning
When retrieval is enough, and when weights need to change.
-
▸ Post
Redis semantic caching for RAG
Cut repeat cost without serving stale answers forever.
-
▸ Post
Circuit breakers for vector DBs
Stop cascading failures when retrieval falls over.
-
▸ Series
RAG in Production series
Ordered path from failure modes through stack and caching.
Ship next
▸ Packages · problem pages · contact
-
▸ Page
Production RAG on Laravel
Problem page: pgvector, hybrid search, evals, cost guardrails.
-
▸ Page
AI & MCP Integration package
The engagement that covers RAG, Claude, MCP, and evals.
-
▸ Page
Hybrid search case study
A RAG pipeline that held up past the demo.
-
▸ Page
Start an AI sprint
One corpus, one answer surface, measure before we scale.
Questions, answered.
Is this hub the same as the production RAG page?
No. This hub is for orientation and reading. The production RAG page is the problem framing for a build engagement on Laravel and pgvector.
Do I need Laravel and pgvector?
Not to learn from the posts. For a build with me, Laravel + PostgreSQL is the default when that is already your stack. A hosted vector DB is fine when the constraints say so.
Will an engagement include evals?
Yes. Shipping without evals is how demo quality dies quietly. See the production RAG page and the AI & MCP package for how that is scoped.
Can RAG sit beside MCP?
Yes. Many teams start with retrieval, then expose tools over MCP. Same package family; we sequence so each piece earns its place.
Ready for RAG that survives evals?
Free discovery call. We pick one corpus and one answer surface, then measure before we scale.