Description
Summary ILRI is a global non-profit research institute working across Africa and Asia. We are building a multi-agent corporate AI architecture on Microsoft 365 and Azure with our in-house IT team, and we need a partner who can engineer it to production standard and leave us able to operate and extend it. Budget and timeframe. Up to USD 50,000 through 31 December 2026; indicatively up to two days per week, with the possibility of extension into 2027. This is a software engineering engagement, not a no-code or prompt-only one. Before applying, please make sure you can evidence: • production Python and maintainable backend engineering; • enterprise API integration with authentication and authorization — Microsoft Graph / Entra ID preferred; • Azure OpenAI deployed on Azure with secure secrets, logging and monitoring; and • RAG with vector stores plus multi-agent orchestration, evaluation, guardrails and human review. Chatbot demos, GPT wrappers and single-framework prototypes do not clear the technical screen. We need a named technical lead and a named backup. You do not need to be an established agency — an outstanding individual with a credible regular collaborator or bench is welcome. Work is task-based, with the level of effort varying by task. We care much more about verified delivery than proposal polish.Finalists complete the same short paid technical exercise, timeboxed to four hours, followed by a live technical interview with our IT team. Full Terms of Reference attached. Separate opportunity: if your strength is AI adoption, change management and ways of working rather than production engineering, see our AI Workflow & Adoption posting instead. SCREENING QUESTIONS 1. Describe one production system you have built that connects an enterprise API (Microsoft Graph or comparable) to an LLM: what it did, what was specifically your work, and a reference that can verify it — repository, live system, or client. 2. Name the technical lead and backup who would do this work, give their current weekly availability, and tell us whether you have delivered together before. 3. What is the weakest part of your fit for this brief? We would rather discuss a real gap than discover it later. 4. How have you used AI to prepare and apply for this role? Be specific — we are genuinely interested, not looking for a particular answer. SUGGESTED SKILLS • Python • Microsoft Azure • Azure OpenAI • Microsoft Graph • Generative AI / AI Agents