Arabic needs context, not translation
Dialect, measurement, and user experience in conversational automation.
Arabic in production is rarely MSA-only. Gulf customers write the way they speak; a bot trained on formal news corpora can sound grammatically correct and still fail the conversation.
We pair dialect-aware design with evaluation: grounded answers, clear handoffs to humans, and model choices that survive cost and compliance reviews — the essays below walk through that stack.
- GPT-4 vs Claude vs Gemini — an objective comparison.
This is not a popularity vote. It is a decision frame: what differentiates each family, where each leads, where each weakens, and how to choose without buying the myth of a single "best" model.
- How the Transformer works — a plain-language guide.
"Attention Is All You Need" changed the industry, but it does not belong in a product review meeting. This is the version for builders: one mechanism called attention, reweighting importance between tokens based on context — without a single equation.
- Why most Arabic AI bots fail.
It is not the model. It is that we train it on Arabic no one actually speaks, then act surprised when no one understands it back.
- Inference vs training for LLMs — who pays for what.
Training might run once (or for many hours) and you pay a cluster bill. Inference runs forever and turns a model into a per-token Opex line. This article separates the two checkbooks so pilot budgets are not mixed with product bills [1].
- What is vLLM — and why production teams use it.
vLLM is an open inference engine for LLMs: scheduling, continuous batching, and KV memory designs such as [PagedAttention](/en/journal/what-is-pagedattention-llm-serving-2026). The point is not a thin API wrapper — it is raising useful throughput under real traffic [1].
- What is LoRA — and how it cuts fine-tuning cost.
When people say fine-tuning, many still picture updating billions of weights in an expensive full pass. LoRA freezes the base and injects a low-rank delta into selected linear paths — often enough to shift behavior on a narrow task without shipping a full weight copy. This article explains the idea without hype, and when savings move from slides to investment [1].
- Five RAG metrics to check before you blame the LLM.
Before you raise model spend or switch vendors, measure retrieval, chunks, and escalation. Most production hallucination starts in documents and indexes — not parameter count.
- What is RAG — and why your company bot answers like a stranger.
A practical guide to Retrieval-Augmented Generation: how your bot reads documents before answering, and why it costs 10× less than fine-tuning.
- What is fine-tuning — and how it differs from prompting.
Half the meetings say "we will tune the model" while they mean "we will rewrite the prompt." The two complement each other — but one changes the text going in, and the other can change the model's weights. That distinction clarifies the decision and saves you from training costs you did not need.
- What is PagedAttention — and what it changed in LLM serving.
Serving bottlenecks were not always raw GPU speed; they were often KV cache waste. PagedAttention changed the equation by treating KV memory as pageable blocks instead of large contiguous reservations, cutting waste and lifting throughput on the same hardware.
- What is a large language model — complete guide for 2026.
This is not a glossary entry. It is the operating calculation behind LLM decisions in 2026: how the model works, where it fails, and how to choose the right deployment path.