Builders
You build the thing that has to survive production.
5 steps · 42 posts
Since your last visit Continue from there →
Step 1
What it is
1 post Step 2
It answers
6 postsThe augmented LLM
Prompt engineering from zero: write a formal letter, not a text message
Zero-shot, few-shot, or chain-of-thought: picking the right technique
From prompt engineering to context engineering
Working with LLM APIs: first calls, tokens, and structured output
How agents remember: memory and knowledge representation
Questions beside this step Klarna went all-AI, then walked it back
Step 3
It acts
6 posts Step 4
It runs the work
8 postsWhen one agent isn't enough
The Coordinator, the Worker, and the Delegator
Event-driven by design: agent teams that don't lose messages
Tracing a request through a multi-agent system
Save first, then publish: a simple rule for not losing work
Cognitive Least Privilege: your agent should know only what it needs
2 more in this step
All of this step → Questions beside this step Framework or platform? Choosing the right abstraction for your agent system
Step 5
You can trust it
21 postsRun it for real
Fine-tune, RAG, or prompt: which one, and what each costs
PagedAttention and continuous batching: how one server answers more users
Stop ranking LLMs, start profiling them
A practical checklist for picking an LLM for your feature
How LLM inference actually works: prefill vs decode
15 more in this step
All of this step → Questions beside this step Off-the-shelf, or adapt your own?