You test a prompt in a web playground. The model parses your request, applies the logic perfectly, and returns a flawless response. You paste that exact string into a production application, and the illusion breaks. The model hallucinates an outdated product catalogue. It forgets the userβs previous message. It outputs conversational text instead of the strict schema your database expects.
The web interface masked the reality of the underlying infrastructure. A base model endpoint is entirely stateless, isolated, and blind. It has no memory of the last turn, no connection to your live data stores, and no tools to verify facts.
A bare prompt is not a system. At this stage of deployment, the working unit is the augmented LLM: the model paired with external retrieval, memory management, and function calling. You take a raw inference engine and wrap it in deterministic scaffolding. You supply state by injecting carefully pruned conversation history. You supply grounding by retrieving relevant internal data immediately before the call. You supply constraints by enforcing rigid output structures.
There is no autonomy at this rung. The system operates strictly within a requestβresponse cycle governed by your application code. You are still driving the orchestration. The majority of the engineering value here does not come from attempting autonomous loops, but from disciplined context engineering β systematically managing exactly what information occupies the modelβs finite context window at the moment of evaluation.
This stage establishes the baseline for everything that follows. Before handing over control flow, you must first reliably extract structured, predictable work from a stateless endpoint.
- Prompt engineering from zero: write a formal letter, not a text message β the discipline of asking precisely.
- Zero-shot, few-shot, or chain-of-thought: picking the right technique β matching the method to the task.
- From prompt engineering to context engineering β the shift from wording to what fills the window.
- Working with LLM APIs: first calls, tokens, and structured output β the mechanics of driving the model in code.
- How agents remember: memory and knowledge representation β giving the stateless engine a past.