Architecture Lab

Educational reference architectures for enterprise data and AI systems. Search the four categories. Illustrative studies, not client deployment claims.

Reference architecture - illustrative, not a client deployment

16 published studies

LLM & Agentic AI · FeaturedHow Enterprise RAG Actually WorksTrace how enterprise documents are indexed with permissions, retrieved for an authenticated caller, checked for grounding and cited, or refused when the evidence is too weak.Intermediate
LLM & Agentic AI · FeaturedFrom Prompt to Action: Anatomy of an AI AgentDistinguish model proposals, runtime enforcement and actual business actions.Intermediate
LLM & Agentic AIMulti-Agent Systems Without the HypeExplain delegation, bounded coordination and when one agent is sufficient.Intermediate
LLM & Agentic AIAn AI Copilot Inside a Business ProcessFollow a support ticket from assisted drafting to an approved system update.Intermediate
ML & MLOpsWhat MLOps Looks Like in the Real WorldSeparate training, release decisions, serving and gated improvement.Intermediate
ML & MLOpsReal-Time Fraud Detection: Models and RulesSeparate the fraud score from the policy that allows, reviews or blocks a transaction.Advanced
ML & MLOpsFrom Demand Forecast to Pricing DecisionExplain why forecasting demand and choosing a price are different problems.Advanced
ML & MLOpsInside a Recommendation EngineShow retrieval, ranking and business filtering as separate stages.Intermediate
Enterprise AI PlatformsAI Inside an Event-Driven Integration PlatformFollow an event through enrichment without duplicating the business action.Advanced
Enterprise AI PlatformsEnterprise AI Platform: The Big PictureSeparate request-time services from platform delivery and governance.Advanced
Adaptation & EvaluationHow to Evaluate an LLM SystemCombine deterministic checks, calibrated judges and human review.Advanced
Enterprise AI PlatformsHybrid AI: What Stays Private?Show that data location, operational control and sovereignty are different dimensions.Advanced
Adaptation & EvaluationInside a Production-Grade Fine-Tuning PipelineDistinguish SFT objectives, full updates, LoRA adapters and QLoRA.Advanced
Adaptation & EvaluationPrompting, RAG or Fine-Tuning?Compare runtime context with weight updates, including a combined approach.Intermediate
Enterprise AI PlatformsSecure AI Beyond the Content FilterExpose identity, data access and tool-execution boundaries.Advanced
Adaptation & EvaluationServing Multiple Models SafelyRoute requests under policy constraints rather than price alone.Advanced