> llms in production
RAG · agents · fine-tuning · LLMOps. Built to run, not to demo.
Four ways I help teams put AI into production — explained in plain language.
Searchable knowledge, in plain language.
Your company's documents, manuals, policies, and product data become an assistant anyone can talk to. People ask questions naturally — "what's our refund policy?", "which clause covers this?" — and get answers grounded in your own content, with citations back to the original source. No more digging through folders.
AI that does the work, not just describes it.
Most AI today only answers questions. Agents go further: they plan, use tools, and complete tasks on your behalf. Think research that pulls from a dozen sources, triage that sorts incoming requests, or workflows that used to require a person clicking through ten different systems.
Custom AI, made for your business.
A general AI model is trained on the entire internet — yours can be trained on the things that actually matter to you: your data, your tone, your specialised tasks. The result is a smaller model that is cheaper to run, faster to respond, and noticeably more accurate at the work you actually do.
Make sure your AI keeps working in production.
Shipping an AI feature is only the start. You also need to see what it is doing, catch when quality drops, and roll back safely when something breaks. LLMOps is the monitoring, evaluation, and release tooling for AI systems — so you spot problems before your users do.
LLMs & agents
LLMOps
ML / training
Infra & MLOps
Data platforms
Cloud
Backend
Hi, I'm Ahmed. I help teams turn AI prototypes into systems that hold up in production.
Over the past 14 years I've been a software engineer, and for the last 8 of them I've worked in AI. I've co-founded and led engineering at an enterprise AI startup, built production natural-language systems used across many languages and countries, and contributed to open-source machine learning tools.
The pattern I keep seeing: AI demos are easy, AI products are hard. The gap between a model that looks impressive in a notebook and a system that serves real users reliably is where most projects stall. That gap is what I help close.
A plain-English guide to retrieval-augmented generation — what it is, what problem it solves, and when it's the right tool (and when it isn't).
Read →Agents are not chatbots. A plain-English explanation of what AI agents actually are, where they help, and where they make things worse.
Read →Shipping an AI feature is the start, not the finish. LLMOps — in plain English — is everything you need so it keeps working once real users find it.
Read →Got an LLM project that needs to ship? Send a short brief.