A cloud based solution architect using AI designs end-to-end solutions on cloud platforms where AI is a first-class capability—not a side experiment. That includes compute/network/data foundations plus model serving, retrieval, agents, and governance.
Nitin Rana combines hospitality/cloud solution architecture experience with practical AI delivery: multi-model chatbots, MCP tool servers, RAG knowledge systems, AI gateways, and optional private Llama stacks.
Cloud baseline: multi-tenant AWS topology, identity, API layers, data stores, CI/CD, HA/DR, and cost controls.
AI plane: model gateway for routing and policy, RAG indexes with ACLs, agent runtimes with least-privilege tools, evaluation and tracing.
Security plane: PCI/GDPR/SOC 2 aligned boundaries so guest and payment data never leak into uncontrolled prompts or logs.
New failure modes: prompt injection, tool abuse, retrieval poisoning, and unpredictable latency/cost. Solution architects must design guardrails, budgets, and fallbacks the same way they design autoscaling and backups.
Hybrid model strategy is often best: private Llama for sensitive workloads; managed cloud models for general tasks—routed through one enterprise AI gateway.
Faster solution delivery with reusable cloud + AI building blocks.
Clear ownership between platform, product, and AI operations teams.
Measurable SLOs for answer quality, tool success rate, latency, and spend.