Skip to content
Zero-ShotTeam

Writing

Notes

Notes on decisions we ran into on real projects. Not commentary on AI in general — what we actually looked at when we had to choose.

  1. 4 min read

    Why AI agent projects fail

    Agents that impress in a demo stall in production. Four observations on tool design, error handling, handing off to a human, and the cost ceiling.

    ajanlarotomasyondeğerlendirme
  2. 3 min read

    Where to start with enterprise AI

    How to choose a first AI project: which tasks qualify, in what order to proceed, and what to expect in the first eighty days.

    kurumsal yapay zekâproje seçimiölçüm
  3. 4 min read

    Writing prompts for production

    Why a prompt that works in a demo breaks in production. Treating prompts like code: structure, versioning, boundaries and measuring every change.

    promptüretimdeğerlendirme
  4. 4 min read

    Why AI makes things up, and how to reduce it

    Why language models state false things confidently. The real cause of hallucination, which countermeasures work, and which do not.

    halüsinasyondoğrulukRAG
  5. 4 min read

    Is it safe to send company data to AI

    What happens when company data reaches a language model: provider terms, training use, data protection law, cross-border transfer and self-hosting.

    veri gizliliğiKVKKkurumsal yapay zekâ
  6. 4 min read

    Why visual inspection gets hard in the field

    Why a vision model that scores 99% in the lab drops on a production line. Field notes on lighting, camera placement, rare defects and data collection.

    görüüretimkalite kontrol
  7. 3 min read

    Making LLM costs predictable

    Why AI spend surprises you at month end, and how to control it: model tiering, prompt caching, output length, and per-request ceilings.

    maliyetLLM altyapısıönbellek
  8. 3 min read

    How to build an evaluation set for an AI project

    How do you prove an AI system works? An evaluation set that starts at fifty cases and carries every decision from a prompt tweak to a model upgrade.

    değerlendirmeajanlartest
  9. 3 min read

    Six common mistakes when building RAG

    Why enterprise retrieval systems demo well and fail in production: chunking, retrieval ranking, citations, permissions, freshness and measurement.

    RAGkurumsal aramaerişim
  10. 4 min read

    Zero-shot, few-shot or fine-tuning?

    Prompting alone, adding examples, or training a model on your own data? The order to try them in, what each really costs, and when to move on.

    ince ayarpromptmodel seçimi