AI Governance and Multi-Tenant Orchestration Intern - Elchai Group

AI Governance and Multi-Tenant Orchestration Intern.

Company: Elchai Group

Location: Headquarters, Dubai — In-person

Schedule: Monday–Friday

Duration: two months

Compensation: Unpaid training program, with a certificate upon completion and the possibility of a paid role based on performance and the company's needs.

Reporting to: Technical Lead of Governance Orchestration, with direct support.

Mission

Learning to build and run agents within our governance orchestration, The Universe Eye Governance Orchestration , respecting the two laws:

• Isolation of each client branch.
• A human approving every important action.

This isn't a tool research role; it's a journey to understand where the controls are connected while the system is running.

What you will learn and do

• Build and configure an isolated client branch, with separate data, roles, and permissions, and understand why one client should never see another's data.
• Work on the real building blocks of orchestration: request routing, narrowly scoped identities for each agent, tool permissions, human approval gates, state mediation, observability, and rollback.
• Build and test the flow where the agent proposes, the system applies the rules, and the human approves before the action starts.
• Learn customer traceability, i.e., logs and audits that show which agent did what, for which customer, and who approved it.
• Understand the behavior of versioned, tracked, auditable, and reversible agents.
• Document clear procedures and SOPs, because school is the factory of procedures.

The final product of the process

At the end of the two months, you deliver a complete customer branch with its six components:

• Customer profile
• Sales lines
• Isolated knowledge base
• Templates
• Compliance rules
• Human gate

Documented, tested, and ready to be cloned.

It's concrete proof of what you've learned.

Requirements

• Basic understanding of how AI agents and automations work.
• Ability to research independently and learn new tools without constant supervision.
• Basics of role-based access control, APIs, terminal, and GitHub, appreciated.
• Strong written English for reports and SOPs, with the ability to explain technical things in a simple way.
• Methodical and obsessive mindset about detail — you don't assume something will work, you try.

Preferred

Basic Python or JavaScript, interest in AI governance, multi-tenant systems, real-world asset tokenization, forensic AI, or data centers, ability to do demos and recordings.