Most AI projects end as a chatbot nobody uses. We do the opposite. We sit with a company, learn how it actually runs, and build the system its people and AI agents work in every day.
We're looking for the engineer who builds and owns that system, from the database up.
What the work looks like
• Design the architecture of platforms where people and AI agents share one source of truth: data model, roles, permissions.
• Build agents that act on real company data, triggered from WhatsApp, email and apps, and not locked to a single model provider.
• Connect what companies already use (ERPs, CRMs, Google Workspace) instead of replacing it.
• Deploy and run everything on servers you set up: security, backups, monitoring, uptime.
• Turn what we build once into reusable blocks we can deploy for the next company.
• Work closely with our design team, and sometimes on site with the teams using what you built.
You'll fit if you
• Have 4+ years building and shipping production software.
• Can plan an architecture that scales, and explain the trade-offs behind it.
• Take security seriously: row-level security, access control, secrets, isolation between users and agents.
• Have deployed and maintained your own servers and databases, not only managed platforms.
• Have built LLM features or agents with tool calling that real users rely on.
• Work fluently in Arabic and English.
• Are based in Lebanon.
Stack
We choose tools to fit the problem, and you'll have a real say in it. Expect Postgres, TypeScript or Python, LLM APIs (Anthropic, OpenAI), and Docker on Linux.
Nice to have: n8n, WhatsApp Cloud API, Odoo or other ERP integrations.
This role is not for you if
Your AI experience is prompting and no-code tools only. We need someone who can build, secure and maintain the server and database underneath, not just the layer on top.
What you get
• Ownership of the architecture from day one, working directly with the founder.
• Real operational problems from companies across industries, not a ticket queue.
• A high-end workstation with an RTX 5090 for local experiments.
• An office in Dbayeh, with flexibility for hybrid days.
How to apply
Apply through this short form (about 10 minutes): https://tally.so/r/J1KEZ7
- Process: short call → short practical test → final meeting.
