← Articles
AM

Aisha Mwangi

Author

Frontend engineer at a Nairobi fintech, building performant React UIs for low-bandwidth markets. Invented the predictive prefetch strategy that cut LCP by 40% on 2G connections.

reactperformancemobiletesting

Articles by Aisha Mwangi(3)

LLM context window management: chunking, compression, and summarisation

Managing what goes into an LLM's context window is as important as choosing the right model—poor context management causes hallucinations, high latency, runaway costs, and silent truncation failures. This article covers chunking strategies, prompt compression, rolling summarisation, and retrieval-based context selection for production LLM applications.

context-windowchunkingcompressionsummarisationperformance

16 July 2026

Aisha Mwangi — ANN Tech — ANN Tech