How do you manage a model's limited context?
Review note: The answer refers to the original author's CV or target job offer — adapt it to your own experience.
Answer out loud or in your head first, then compare.
Answer
I pass the information needed for the current step, not everything I have. A long context can cost more and make it harder for the model to find what matters. I also have to leave room for the answer and further tool results.
When working in a repo, I first give the goal, the rules and pointers to the relevant files. I read in details when they're needed. I store important decisions and state outside the conversation, so I can recreate them. A summary saves space but can lose a condition, e.g. a ban on changing the API. So key constraints should be explicit and checkable. A bigger window helps, but it doesn't release you from choosing information.
Why it matters: Directly concerns the context and harness from the offer.
Possible follow-ups: What must survive shortening the history or starting a new session?
How did it go? Be honest — “Not yet” puts the question on your review list.
Start with
- F-P01 · What is an LLM? — LLM fundamentals and AI engineering
- F-P03 · What are a token and a context window? — LLM fundamentals and AI engineering
Further reading
- Anthropic: Effective context engineering for AI agents(opens in a new tab)29.09.2025. Choosing context, limits, and references instead of blindly growing the input.