Nuno Salvação

• • Designing and maintaining E2E architecture from ingestion to observability, covering data pipelines, orchestration, and infrastructure automation.

• • Developing and validating multi-stage retrieval (BM25 + dense + rerank) achieving >95% citation accuracy, measured with RAGAS harnesses and Langfuse traces.

• • Delivering LLM orchestration with caching, rate limiting, and FinOps tracking.

• • Delivering enablement assets (runbooks, ADRs, DPIA summaries, FinOps baseline) for repeatable adoption.

• • Implementing zero-trust perimeter access and privacy-by-design ingestion.

• • Contributing to cross-domain solution design (data, AI compliance, automation) aligning guardrails and risk controls with GDPR Art. 25 and AI Act readiness.

CET Level 5 (EQF 5) programme in IT Systems & Network Management focused on operational delivery: network design and security (routing/switching, VLAN segmentation, DHCP, access control), Windows Server and Linux administration, and systematic hardening and troubleshooting. Complemented by cloud infrastructure (Azure/AWS/GCP/DigitalOcean), Docker/Compose for service packaging, and automation via Python/PowerShell. Current projects explore practical AI integration for ops workflows (e.g., assisted diagnostics, documentation and multi-model experimentation) with attention to cost and governance.

Led consultative engagements for IT infrastructure, cloud & IoT solutions. Collaborated directly with engineering teams on integration optimization and solution architecture. Technical value: Deep exposure to enterprise AI/automation requirements, building foundation for current hands-on development work.