# Home URL: https://donedeploy.com/ [!AI · November 2, 2026CI/CD for Models and PromptsEvaluation gates, model registries, canary releases and rollback: applying CI/CD discipline to prompts and models in LLM applications.](https://donedeploy.com/cicd-for-llms/)[!AI · October 23, 2026Fine-Tuning or Prompting? A Decision Guide for Small TeamsWhen prompting, RAG or fine-tuning is the right way to adapt a language model, and what each one costs a small engineering team.](https://donedeploy.com/fine-tuning-or-prompting/) [All articles »](https://donedeploy.com/blog/) Questions ## Frequently asked What does DoneDeploy do? We design, build and run cloud infrastructure for software companies, with AI in the loop: AI-assisted service design, infrastructure as code on AWS, Google Cloud and Azure, Kubernetes, CI/CD, AI observability, and AI that automates business processes. How do you use AI in infrastructure and cloud work? AI drafts and reviews Terraform and Kubernetes changes, reads metrics and logs to spot anomalies and group alerts into incidents, recommends cost and capacity changes from real usage, and runs automated runbooks. Engineers stay in charge of every change that reaches production. What is AI observability? Observability with AI on top: the usual metrics, logs and traces, plus models that detect unusual behaviour, cut alert noise, suggest likely root causes and summarise incidents in plain language. For LLM applications we also track tokens, latency, cost and answer quality. Can you run open models like Llama, Mistral or Qwen on our own servers? Yes. We deploy open models with vLLM, Ollama or TGI on Kubernetes GPU nodes in your cloud account or on your own hardware, so data never leaves your environment, and route to hosted models such as GPT, Claude or Gemini where they are the better fit. How do we start using AI in our company? Start with the work, not the tool: we look at where your team loses time, pick one or two processes where AI clearly helps, run a short pilot on your own data with clear success measures, and set simple rules on data and tools before rolling it out. Our Getting started with AI service covers exactly this. What should we automate first? The work that is frequent, repetitive and rule-based: deployments and environment setup on the engineering side, and on the business side things like triaging requests, extracting data from documents and preparing reports. We map the process first, remove the steps that should not exist, then automate what is left. How can AI improve our business processes? AI is good at reading, sorting, summarising and drafting: support requests, documents, emails, reports and logs. Put into a process with people checking the important decisions, it removes hours of manual work while keeping quality and an audit trail. We measure the process before and after so the improvement is real. Do you work with startups and small teams? Yes, most of our clients are small and mid-size teams. We work on fixed-price packages with a free first consultation, so you know the cost before work starts. Where are you based? DoneDeploy is based in Ireland and works with teams across Ireland, the UK and Europe, remotely.