Unpaid Internship AI Governance and Orchestration

**AI Governance and Orchestration Intern** **Elchai Group | Dubai, UAE | On-site | Internship and School Pathway** **About the Role** **We are looking for an AI Governance and Orchestration Intern** who is interested in understanding how AI agents can operate safely within clear permissions, isolated workspaces and human approval processes. **This role is completely unpaid, on-site, entry-level internship designed for students, recent graduates and early-career applicants seeking practical experience in AI Governance and Orchestration.** In this role, you will work on small, supervised components of **The Universe Eye Governance Orchestration** , helping test workflows, document controls, identify limitations and create repeatable procedures. This is a practical learning opportunity for students, recent graduates and early-career applicants interested in AI agents, workflow automation, access control, testing and technical documentation. **Programme Details** Location: Office 2008, Jumeirah Business Centre 1, Cluster G, Jumeirah Lakes Towers (JLT), Dubai, UAE Workplace: Fully on-site Daily Hours: 10:00 AM to 4:00 PM, Dubai time Working Days: Monday to Friday Break: 1:00 PM to 1:30 PM lunch/rest break Initial Supervised Stage: Three days Internship: Eight calendar weeks Compensation: The initial stage and school stage are unpaid, with an AED 0 stipend. Participants cover routine commuting, meals and accommodation. Elchai provides the approved tools needed for assigned training. **What You Will Learn and Do** **As part of the programme, you will:** * Map small workflows from request through policy checks, approval, action and audit evidence. * Configure or test supervised demonstration workspaces using approved identities and permissions. * Check that one customer, venture or function cannot access another branch's data or tools. * Test approval scope, changed versions, denied access, interrupted execution and safe recovery. * Record reproducible results, limitations and defects for the technical team. * Produce procedures and handovers that another trained person can repeat. * Work with practical checks, documented evidence and feedback to understand how governance controls operate in AI workflows. **The Two Core Rules** **Two rules guide the work:** **1. Keep customer, venture and function branches isolated.** **2. Obtain specific human approval before important actions.** **Work takes place in approved training or demonstration environments. The technical team controls repository access, permissions, security policy, merges and deployments.** **The intern will not activate a customer branch or approve their own consequential action.** **What You Need at Entry** **You should be able to:** * Explain a basic automation or AI-agent workflow in clear English. * Reason about who may read data, use a tool and approve an action. * Check a result and describe a failure case before considering work complete. * Follow a technical brief, document your steps and ask for missing information. * Explain the work you submit, including any permitted AI assistance. **Useful Technical Skills** **The following are useful but not mandatory** : * Basic Python or JavaScript * APIs * Git * Terminal usage * Role-based access control * Workflow automation * AI agents * Technical documentation * Testing Academic and personal automation projects count as relevant experience. You do **not** need production architecture experience to enter the school. Students, recent graduates and other early-career applicants may apply, subject to appropriate training or work authorisation. **What Progress Looks Like** **During your first three supervised days, you will:** * Explain the two core rules. * Test a bounded workflow. * Document its limits. By school completion, your evidence should include: * A scoped demonstration branch * Isolation and approval tests * Audit evidence * A recovery checklist * A repeatable operating procedure ready for human review We assess: * Delivery * Quality * Product and rule understanding * Reliability * Communication **Example Practical Task** **You may be asked to review fictional AI-agent actions and determine whether they should:** * Proceed within the supplied scope * Be stopped * Be referred for specific authorised review You will be expected to explain your evidence and design tests that demonstrate proper isolation and approval controls. Coding is optional. A clear table and test plan are sufficient. **Certificate** **A school completion certificate is proposed to be issued by Elchai World FZCO under the Elchai Group brand** after: * Completion of the eight-week school * Completion of required work and evidence * A score of at least 7.5/10 * Written human sign-off The proposed certificate issue deadline is **ten staffed working days after the required sign-off** . The certificate represents internal training and does not make an academic or government accreditation claim. **AI Use and Human Review** **AI may be used as part of screening and structured interviews. Candidates will be informed about AI use and any recording before the session.** **Final admission, rejection and progression decisions are made by a named human decision-maker** .