Build the decision layer with evidence.
Jev, language models and ordinary code have different jobs. Explore contracts, architecture, calibration, Java implementation and deployment boundaries with inspectable examples.
Jev, LLMs and SLMs: choose by the decision contract
A useful model comparison begins with the output the application needs. A support workflow might need a paragraph explaining a problem, a choice of support…
Guide 02Design a typed decision contract before selecting a model
When a model controls a branch in software, the branch needs a contract. The contract describes the evidence, permitted answers, uncertainty behavior and…
Guide 03A hybrid agent architecture that keeps decisions separate from effects
A practical hybrid system can use a language model to draft, a decision model to classify, and ordinary code to enforce permissions. The useful separation is…
Guide 04Probability, confidence and correctness are different measurements
A decision response can contain a selected option, a probability distribution and a confidence value. Those fields answer different questions. The selected…
Guide 05Build and test a bounded Jev adapter in Java
A Java application does not need to represent every provider feature to make one useful, testable decision. Start with a narrow adapter for a Choice question.…
Guide 06Abuse and fraud prevention needs controls outside the model
An abuse-and-fraud-prevention system should separate suspicious-content judgment from permission enforcement. A model can help prioritize ambiguous evidence. It…
Guide 07Put typed decisions inside ADK workflows without moving policy into prompts
An agent framework coordinates work; a decision model supplies one kind of judgment. Keeping those roles separate makes the system easier to test and replace.…
Guide 08Route between rules, local SLMs and hosted models using measured constraints
A model router is useful only if it improves an application objective. Sending simple requests to a cheaper model sounds reasonable, but the routing call…
Guide 09Use typed decisions to review RAG evidence without inventing certainty
Retrieval-augmented generation has at least three distinct failure points: retrieving the wrong evidence, generating an unsupported claim, and presenting a…
Guide 10Evaluate the decision model and the application as separate systems
A thousand passing tests can establish useful software properties without establishing model intelligence. Conversely, a strong benchmark score can coexist with…
Guide 11Prepare decision data before considering a local-model fine-tune
Fine-tuning is not the first remedy for an unclear decision contract. If reviewers disagree about what the labels mean, training can reproduce that ambiguity…
Guide 12Design a GCP deployment for decisions, approvals and realtime progress
Moving a local agent into a hosted service introduces identity, concurrency, persistence and failure recovery. A process that reads a local environment variable…