Cross Stream Innovation Architect for GAVI
Location : Geneva
Fixed Term or Permanent
Hybrid – 2 days in Geneva office
The purpose of this Cross Stream Innovation Architect is to lead architecture and technical strategy across Innovation streams: AI Products Delivery, Innovative Software Engineering (Industrialized vibecoding), and Citizen Development Enablement (Center of Expertise), ensuring secure, scalable, and measurable solutions by setting standards, designing reference architectures, and embedding Responsible AI and platforms governance.
- Specific duties and responsibilities include:
- Define and communicate the architecture vision and standards for AI, application, data, integration, and security across all streams.
- Make tenant-level AI decisions (tools selection, data residency, safety posture).
- Develop and maintain an enterprise LLM playbook defining model selection criteria, usage guidelines, cost and consumption monitoring, and governance rules for dynamic model switching
- Author reference blueprints for Conversational AI (Copilot Studio), Agentic orchestration (MCP/A2A, tool/function calling), and RAG pipelines (indexing, evaluation harnesses, observability).
- Define evaluation gates (quality, safety, performance) and standardize prompt/grounding patterns and telemetry.
- Translate business needs into end-to-end solution designs (integration, data contracts, release strategy).
- Guide decomposition into services/agents and ensure reusability and alignment with platform roadmaps.
- Support delivery teams (AI engineers, Vibe-code engineers, Power Platform engineers) during build/run.
- Architect Power Platform Core (environments, DLP, connectors, ALM, identity & access).
- Co-design the Center of Excellence operating model for Citizen Development initiative: onboarding, safe zones, quality checks, solution transition to IT.
- Define maker-ready templates/components and guidance for AI-enabled apps/flows/copilots.
- Implement guardrails and AI risks mitigations: content filters, grounding policies, jailbreak mitigation, bias/fairness checks.
- Define incident playbooks (red-teaming cadence, rollback criteria, human-in-the-loop fallback).
- Embed observability (tracing, feedback, drift monitoring) and SLOs (latency, quality, cost).
- Mentor engineers and makers; drive a culture of framework-first delivery (BMAD), documentation, and reuse.
- Comms/Change for stakeholder engagement and interaction; present architecture and results to leadership.
- Qualifications and work experience required / Team Structure:
- Architecture Leadership:
- 5–7+ years in enterprise/solution/AI architecture, with hands-on delivery of production AI or automation systems.
- AI & ML Systems:
- Deep experience with Azure AI/Azure OpenAI, Copilot Studio, RAG pipelines, vector stores, prompt engineering, and evaluation harnesses.
- Application & Integration Architecture:
- Strong background in cloud-native (Azure), APIs, events, security, and integration patterns.
- Power Platform & Citizen Development:
- Proven platform governance for Power Platform (Dataverse, DLP, environment strategy, ALM), and experience enabling citizen development at scale.
- Responsible AI:
- Demonstrated implementation of Responsible AI controls (safety, monitoring, red/blue teaming, bias/fairness).
- Stakeholder Management:
- Excellent communication and influence skills across product, engineering, security, and business teams.
Preferred/Bonus
- Experience with agentic frameworks (BMAD or similar), MCP/A2A integrations, and human-in-the-loop workflows.
- Observability for LLM systems (feedback loops, eval suites, prompt/response telemetry).
Relevant certifications (Azure Architect, Power Platform Architect).