Rome, Latium, Italy
Enterprise Solution Architect | AI, Cloud, and IT Infrastructure Expert With over 20 years in IT, I help enterprises turn AI from proof-of-concept into production-grade infrastructure — on-premise, governable, and sized to the real workload. I translate complex technical requirements into solutions with verifiable numbers behind every choice. 🌐 Key Expertise AI & LLM Infrastructure: Production RAG systems and LLM serving (vLLM, Ollama) on multi-GPU clusters, vector databases (Qdrant), document ingestion with OCR, and Vision-Language models. GPU & HPC: Design and sizing of GPU platforms (NVIDIA H200/H100/A100), high-performance networking (RoCE/RDMA), NCCL performance validation, and Kubernetes/OpenShift orchestration. Capacity Planning: Deterministic hardware-sizing for on-premise LLM deployments, preventing costly over-provisioning through rigorous KV cache and memory modeling. Enterprise Infrastructure: Hands-on across Dell PowerEdge, enterprise storage, VMware/ESXi, SAN/Fibre Channel, and data-sovereignty-driven architectures. 🚀 Notable Projects Designed and deployed a high-performance AI platform for INPS, using PowerEdge GPU nodes and Kubernetes for production inference. Built a production RAG system for scientific-medical literature on an NVIDIA H200 cluster, with measurable gains in retrieval efficiency and accuracy, in partnership with CNR (Italian National Research Council). Led a performance-validation engagement on a customer's H200/OpenShift cluster, diagnosing a structural networking bottleneck and proposing the corrected design, backed by NCCL benchmarks and bilingual documentation. Corrected a GPU proposal with an 8× memory-sizing error, avoiding major over-provisioning by redesigning onto a more efficient, lower-cost platform. 🎓 Certifications Microsoft Certified Azure Administrator Associate | VMware Certified Professional | Cisco CCNA 2024 | Cloud Architect (EMC) 📩 Beyond technical depth, I focus on pre-sales engineering — shaping solutions around real client needs, validating them with proofs-of-concept on real datasets, and enabling teams through technical training. I'm driven by problem-solving and by turning technology into tangible business value