Archive | Prompt/Deploy

Prompt/Deploy
Archive
Page 1

Archive

production-ragproduction-rag

The Chunking Problem: Size, Overlap, and How to Measure What Works

How to think through chunking strategy and build the eval loop that tells you if it's working.

Hou C.

production-ragproduction-rag

Incremental Re-indexing and the Embedding Pipeline Nobody Talks About

Embedding once is easy. Here's the production engineering behind keeping your RAG index correct, fresh, and survivable when you need to change your model or chunking strategy

Hou C.

production-ragproduction-rag

How I'd Choose a Vector Database

The tutorial picks one for you. Here's the decision logic for choosing a vector database in production — so you can eliminate the wrong options before they become expensive ones.

Hou C.

production-ragproduction-rag

When RAG Is the Wrong Answer (And What to Reach For)

You've already shipped RAG. Now some use cases are failing structurally. A four-gate decision tree for knowing when to fix the pipeline vs. when to reach for something else.

Hou C.

Mental Models for Production AIMental Models for Production AI

When Should I Retrain or Tune?

The decision framework for updates. An interactive guide to knowing when to tune prompts vs retrain models vs investigate deeper.

Hou C.

Mental Models for Production AIMental Models for Production AI

Systems That Don't Learn, Decay

Closing the loop. Why continuous learning isn't optional—systems that don't improve, degrade.

Hou C.

Mental Models for Production AIMental Models for Production AI

Silent Failures Make Me Most Nervous

The degradation timeline. How AI systems fail slowly without anyone noticing—and how to catch it.

Hou C.

Mental Models for Production AIMental Models for Production AI

When Do I Need Human Review?

Where humans belong in the loop. An interactive framework for deciding when and how to involve human reviewers.

Hou C.

Mental Models for Production AIMental Models for Production AI

How Do I Know It's Failing?

Detection before it's too late. An interactive guide to setting up failure detection for AI systems.

Hou C.

Mental Models for Production AIMental Models for Production AI

The AI Failure Cascade

Every layer is a potential failure point. Maps the path from user request to user impact and where things can break.

Hou C.

Mental Models for Production AIMental Models for Production AI

Build Discipline vs Run Discipline

Two different mindsets required for AI features. What it takes to build well vs what it takes to run well—they're not the same.

Hou C.

Mental Models for Production AIMental Models for Production AI

Build vs Buy: Vendor Strategy for AI Features

When to use APIs vs build your own. An interactive guide to the build-vs-buy decision for AI capabilities.

Hou C.