Most portfolio guides are written for people chasing full-time jobs. This one is written for AI engineers who want to get paid — contract by contract, client by client. There are thousands of articles telling you to "add 3-5 projects to your GitHub" and "include a live demo." You've read them. So has every other ML engineer applying for the same contract you want. Here's what those articles never say: contract clients don't evaluate you the way employers do. A hiring manager at a big company wants proof you won't break things. A contract client wants proof you'll deliver business outcomes — fast, without hand-holding, and without a six-month onboarding. That's a completely different bar. And almost nobody's portfolio clears it. This checklist fixes that. First, Understand What Contract Clients Actually Screen For Before we get into what to put in your portfolio, you need to understand the mindset of the person reviewing it.…