AI staff augmentation: a senior engineer you direct

The difference is who directs the work, and HorizonLux offers both models. In staff augmentation, the provider employs the engineer and your lead decides what gets built. In managed services, the provider directs the work inside an agreed scope and delivers a result. This page is staff augmentation. The managed route is on our Hire an AI engineer page: a paid pilot on one agreed deliverable, a weekly demo, then a handover with docs, runbooks and training.

On paper, yes: your contract is with HorizonLux, which employs the engineer. In practice, nothing is handed off to a separate team. Your lead directs the day-to-day work, the engineer joins your stand-ups and code review, and every commit lands in your repo. That is the line between staff augmentation and outsourcing a whole project, where the provider decides how the work gets done.

A HorizonLux engineer joins your Slack and repo within five working days of the free call. There is no job ad, shortlist or notice period, because the engineer already works on our senior team. In those days we match someone to your stack and your hours, you meet them, and access to your Slack, repo and board is set up. A good week one: a codebase walkthrough, then an opening pull request on the pilot deliverable, through your review.

HorizonLux agrees the price on the free call, before any work starts, and doesn't publish rates. It is scoped on how long you need the engineer and what they will work on, and billed CONFIRM: billing unit. The paid pilot is priced on its own, so your opening commitment is one deliverable, not a long contract. You get the price, the billing basis and the length in writing before the engineer joins.

You own the code, the IP and the docs, and HorizonLux keeps all of it in your repo from day one. We sign an NDA on request. The engineer gets access only to the repos and tools the work needs, in accounts you control, so you grant that access and remove it when you choose. Commits go to your repo, not a personal account or a side repo. AI coding tools are used only where your policy allows.

HorizonLux lets you stop after the paid pilot if the work doesn't land, and you keep what was built. After that, notice and replacement terms are written into the contract before the engineer starts, so you know in advance what happens if they leave, fall ill or stop fitting the work. The code, history and docs are already in your repo, so whoever picks the work up next starts from there. We don't promise outcomes; we make the exit plain.

Not through this service: HorizonLux employs the engineer for the whole engagement and doesn't recruit onto clients' payrolls. What we do is work alongside your own search. The engineer ships while you hire, then hands over to your new hire with docs, runbooks and training, so the new person starts from working code and written notes, not a blank page.

Yes, when your stack overlaps with HorizonLux's. Our engineers build AI features and agents on OpenAI and Claude models, and the product around them with React, Next.js, React Native and Expo, NestJS, Postgres, Prisma, Redis and AWS S3. They use Claude Code, Cursor, Codex, n8n and Supabase daily, and work remotely in English, in any timezone. If your stack is outside this list, Seif says so on the call.