Field notes

Clear thinking behind the finished product.

Practical notes on frontend, backend, mobile, realtime systems and production AI.

Writing library

Find the depth, topic and format you need.

4 / 4 articles
Category
Refine results
AI EngineeringArticleFeaturedIntermediateEnglish

Context engineering is replacing prompt engineering for serious AI products

A strong prompt cannot rescue weak context. The harder engineering problem is deciding what an agent should know, retrieve, remember and forget at every step.

8 min read
AI EngineeringArticleFeaturedIntermediateEnglish

Agentic RAG: when retrieval needs reasoning, not another similarity search

Basic RAG retrieves once and hopes for the best. Agentic retrieval can decide what to search, reformulate weak queries and stop when the evidence is actually sufficient.

9 min read
ArchitectureArticleFeaturedIntermediateEnglish

Most applications do not need a team of AI agents

Multi-agent architectures are powerful for parallel work and context isolation, but splitting one solvable task into fake digital employees often adds coordination bugs.

8 min read
EngineeringArticleIntermediateEnglish

Moving a product from REST to GraphQL—without the drama

A practical field guide to reducing over-fetching while keeping a shipping product stable.

6 min read