Loading…
Architecture documents from real cases — guides/deep dives, ADRs, design docs, post-mortem analyses and teardowns — with my reading as a solutions architect.
A practical, opinionated tutorial on using SAP LeanIX as a transformation accelerator in the architect's daily work — covering inventory (Fact Sheets, meta model, quality), diagrams (types, when to use, live vs static), and reports (question → data → decision). Uses Hohpe's Architecture Elevator as the connecting thread and a fictional insurer/bank scenario to make every concept concrete.
The bottleneck in AI agents has shifted from the prompt sentence to the loop design: trigger, topology, verifier, and stop rules. This guide teaches loop engineering fundamentals — from the base ReAct loop to a pattern catalogue — with real examples from my own system and a step-by-step guide for engineering any loop safely.
Choosing between Lambda and ECS is not about preference — it is about matching the unit of scale to the load pattern. This guide covers every Lambda type (including MicroVMs and Managed Instances) and ECS type (Fargate, EC2, Managed Instances), the decision framework I use in practice, and the real impact on engineering, business, and customer experience.
What was advanced in 2025 became default in 2026: the bottleneck moved past the prompt, agents went to managed production, MCP standardized tooling, security became layered, and token FinOps entered the roadmap. This playbook maps each shift to concrete actions — for people building systems, not demos.
A technical lesson for developers and architects who hear 'AI agent' every day but want to truly understand what differentiates an agent from a plain LLM, a fixed pipeline, or simple RAG. We cover the full anatomy — model, tools, memory, planner — and the ReAct loop step by step with a concrete example. No hype; real trade-offs.
The second lesson in the series maps the full catalog of AI agent architecture patterns: from single-agent loops (ReAct, Reflexion, Plan-and-Execute) to multi-agent orchestration, covering memory as an architecture decision, guardrails, and human-in-the-loop. The goal is to give the architect a precise vocabulary to choose — and justify — the right pattern for each problem, without falling into classic anti-patterns.
The third and final part of the series descends the elevator to the technical floor: how to run an AI agent in production on AWS using Amazon Bedrock AgentCore. We cover the full component map (Runtime, Gateway, Memory, Identity, Observability), model selection by cost/latency/reasoning, security with guardrails and session isolation, and FinOps to keep real costs under a few dollars per month.
The new iterative refinement and ambiguity reduction workflows in Bedrock Guardrails Automated Reasoning checks reduce the manual burden of maintaining formal policies — but introduce non-trivial architectural decisions around governance, policy lifecycle, and CI/CD pipeline integration. In this ADR, I analyze the context, options considered, and real consequences of this decision in regulated financial environments.
Architecture decision on how to safely execute AI-agent-generated or user-submitted code in a multi-tenant environment, comparing Lambda MicroVMs (AWS launch June 2026), ephemeral containers with gVisor/Firecracker on ECS/EKS, and standard Lambda. The recommended decision is per-session Lambda MicroVMs orchestrated by Bedrock AgentCore, providing VM-level isolation without the operational overhead of managing your own virtualization stack.