How the Tensor9 compiler resolves service replacements when the values it needs aren't primitives — backward data flow over a typed graph, symbolic conditions, and count-gated template specialization. Every gate resolves at Terraform plan time; no runtime magic.
How Tensor9 models services like PostgreSQL, Kafka, and RabbitMQ; transforms between cloud and on-prem manifestations; and deals with the differences in configuration spaces between source and target.
A deep dive into STIR (Stack Intermediate Representation), the projection model for cross-trust-boundary deployment, and how we transform infrastructure from one form to another.