Train models to understand cache read and write flows, state dependencies, and consistency behavior through protocol source code and structured reasoning targets.

Follow the complete tool-use workflow for meal delivery support, from order verification to replacement or refund.

Train models to recognize CRM chat requests across contact management, sales follow-up, activity logging, and reporting workflows.

Verify whether microsegmentation policies are correctly reconciled and enforced, even when controller execution encounters complex failure states.

Explore the source-level request flows behind admission controllers and webhooks, including policy execution, security checks, and response generation.

Examine comparative ad claims with structured evidence, qualifying conditions, and traceable fact-check labels.

Trace cache-change dependencies through source code to identify invalidation scope and potential dirty-data risks.

Train models to detect conflicts across sources and judge whether citations adequately support synthetic social advertising claims.

Train and evaluate test generation and repair with practical Python examples for artifact digests, checksums, and signature verification.

Infer task dependencies and execution order from data flows in source code to build code models that better understand scheduling logic.

Explore end-to-end cases for verifying real-time messaging code and diagnosing concurrency defects, from failure analysis to suggested fixes.

Explore signaling workflows from meeting creation through termination, with test and defect samples for state-transition verification and root-cause analysis.

Train specialized geology workflow code generation, from resource estimation specifications to implementation.

Use dependency reasoning examples for chassis control software to trace how localized code changes may affect callers, downstream logic, and vehicle control functions.

Evaluate how well models understand source code management permissions with API cases covering authorization decisions, rule conflicts, and violation localization.

Combine timing reports with design context to train models to interpret critical paths and diagnose timing violations.

Evaluate specialized code verification and debugging with transmission diagnostic state machine code, transition constraints, and annotated defect causes.

Trace cross-workflow waiting chains and trigger conditions in scheduler source code, and understand what happens when dependencies are released.

Train models to repair DevOps platform unit tests using source code and flawed tests as context.

Evaluate how well models identify true food product relationships across multilingual records and highly similar hard negatives.