AI Prompt Engineer – Professional Workflows

We work with professionals — people with real jobs, real deadlines, and zero interest in becoming AI specialists — and we help them get consistent, genuinely useful output from AI on a regular workday. Not through shortcuts. Through structure. Through understanding what a well-built prompt actually is, how it’s different from a bad one, and why that difference compounds across hundreds of interactions every week.

We’re a small, deliberately focused team. We build slowly and carefully, and we’re about to launch what we’ve been building for a long time. We’re looking for a Prompt Engineer who wants to work somewhere that takes this work as seriously as they do.

What this role actually is

Most prompt engineering roles ask you to write better prompts for a product someone else designed. This one is different.

At Solen, the prompt systems you build are what the team runs on — content creation, coaching workflows, internal research, student support, sales outreach, data analysis. You’re not supporting someone else’s infrastructure. You’re designing the one we all use. That means your choices about architecture, naming, chaining, and documentation shape how a company that thinks about this for a living actually thinks about it internally.

That’s an unusual amount of real intellectual weight for a contract role. It’s also what makes the work interesting.

What you’ll be building

You’ll start by auditing how our team currently uses AI and identifying every place where a better-structured prompt system would save time or improve output quality. From there, you’ll build and maintain a structured internal prompt library organized by use case — not a folder of text files, but a genuinely navigable system with documentation non-technical team members can actually use.

You’ll design multi-step prompt chains for repeatable workflows: the ones we use every day that currently take longer than they should.

You’ll run regular quality evaluations on output, iterate based on what you find, and keep a running log of what changed and why. And as our understanding of what works deepens, you’ll help translate that into frameworks the wider team can apply independently.

The specific use cases span content, coaching, sales, curriculum design, and analysis. Some will be highly structured. Some will be messier. All of them will require genuine judgment about when a prompt architecture is working and when it only looks like it is.

What we’re looking for

We’re not looking for someone with a LinkedIn headline that says “Prompt Engineer.” We’re looking for someone who has actually built systems — libraries, chains, structured workflows — and can point to concrete differences in output quality that resulted from their architecture choices versus someone else’s intuitions.

You should be genuinely fluent in at least one frontier model (Claude is what we use most; familiarity is a strong advantage). You should understand the difference between role prompting, chain-of-thought, constrained output generation, and multi-step decomposition — not as vocabulary to use in an interview, but as tools you reach for instinctively depending on the problem. You should be able to explain why a prompt works or doesn’t in plain language, because the people you’ll work with aren’t engineers and the systems you build need to be understood by people who aren’t you.

Strong documentation instincts matter here — probably more than in most roles. A prompt system no one else can maintain or improve is a liability, not an asset. If you write documentation the way you’d want someone to write it for you, that’s the right instinct.

Background in education, professional development, or any field where you’ve had to explain a technical process to non-technical people is genuinely useful, not just a bonus.

What working here is like

We’re a remote team that moves quickly, disagrees openly, and changes direction when we’re wrong without making it a production. We have high standards for output quality — partly because of the work we do, and partly because people with high standards for their own work tend to attract more people with the same. You’d be working closely with practitioners who have spent considerable time studying how professionals actually use AI versus how they think they use it. That’s a specific kind of intellectual environment, and it tends to make the work sharper for everyone in it.

This is a contract role with genuine path-to-full…