AI infrastructure & software
Articles on choosing and running AI software: licences and contracts, private AI infrastructure, data residency, costs and security.
13 articlesClear filters
How to host an LLM privately in a UAE data centerHow to host an LLM privately in the UAE: a buildable blueprint with in-country GPU capacity, Kubernetes MIG isolation, vLLM serving, an OpenAI-compatible gateway and a PDPL control map.The best Arabic LLM for your enterprise is the one that passes your eval, not the leaderboardHow to choose between Falcon-H1-Arabic, ALLaM, Fanar 2 and Jais 2 with a week of measurable testing: your own 300-item eval set and the tokenization arithmetic.Confidential computing AI inference in the UAE: GPU TEEsResidency in a UAE data centre does not stop the host reading your prompts. Confidential computing on GPUs can. What to demand, what it costs, and what it does not fix.vLLM vs SGLang vs TensorRT-LLM: how to choose a self-hosted inference engine in 2026Compare vLLM, SGLang, and TensorRT-LLM for self-hosted LLM serving: traffic shape, audited MLPerf results, operational costs, and benchmarking methods.How to evaluate AI software before buying: a hotel guideA practical framework for how to evaluate AI software before buying: score 50 past guest enquiries on accuracy, source traceability and strict pass gates.Why is my AI API bill so high? Where tokens goToken prices fell, but agents multiplied token usage by 10 to 30 times. Here is where the spend goes and the controls that bring invoices back down.
Archive


