Real production systems, names scrubbed. The architecture, the trade-offs, and the numbers are exact.
Nothing here is a toy or a tutorial. Every writeup is a system that actually ran — under compliance gates, latency budgets, and constraints you don’t get to wish away.
Each one carries a C4 container diagram, a Y-statement that commits to the decision, a fitness function that enforces it, and at least one real number — p95, recall@k, cost. No hand-waving.
Building an LLM agent for tax-compliance document review under LGPD, internal audit gates, and controlled deployment windows. Pattern: retrieval pipeline + citation extractor + auditable decision log + human-review gate. Stack: Python + Azure OpenAI + Postgres + Terraform.