Requirements

ML Execution		status: valid security: NO safety: QM
reqtype: Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__enablement satisfied by: feat__ai_platform fulfilled by: logic_arc_int__ai_platform__inference_if
The platform shall support the execution of traditional ML models on embedded automotive hardware.
feat_req__ai_platform__workloads_execution		Feature Requirement

ASIL-B ML Inference		status: valid security: NO safety: ASIL_B
reqtype: Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__safety_critical satisfied by: feat__ai_platform fulfilled by: logic_arc_int__ai_platform__inference_if
The platform shall support deployment of safety-certified inference backends and abstraction layers for ASIL-B use cases.
feat_req__ai_platform__safety_backends		Feature Requirement

Efficient Embedded Runtime		status: valid security: NO safety: QM
reqtype: Non-Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__runtime_efficiency satisfied by: feat__ai_platform
The platform shall optimize runtime performance and memory usage to meet the constraints of automotive edge hardware.
feat_req__ai_platform__embedded_efficiency		Feature Requirement

Cross-OS Portability		status: valid security: NO safety: QM
reqtype: Non-Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__platform_portability satisfied by: feat__ai_platform
The platform shall support portable AI components that run on both QNX (ASIL-B) and Linux environments.
feat_req__ai_platform__qnx_linux_support		Feature Requirement

Verified Model Execution		status: valid security: YES safety: QM
reqtype: Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__model_security satisfied by: feat__ai_platform
The platform shall ensure model artifacts are verified via cryptographic signatures before execution, and model loading is restricted to trusted paths.
feat_req__ai_platform__model_verification		Feature Requirement

Static Backend Selection		status: valid security: NO safety: ASIL_B
reqtype: Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__runtime_determinism satisfied by: feat__ai_platform fulfilled by: logic_arc_int__ai_platform__inference_if
The platform shall select inference backends statically at build time to ensure deterministic runtime behavior.
feat_req__ai_platform__static_backend		Feature Requirement

AI Agent Resource Budgets		status: valid security: NO safety: QM
reqtype: Non-Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__agent_execution satisfied by: feat__ai_platform
The platform shall schedule AI workloads via the existing execution management mechanisms with configurable resource budgets per criticality domain.
feat_req__ai_platform__agent_execution		Feature Requirement

Model Artifact Lifecycle Management		status: valid security: YES safety: QM
reqtype: Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__model_lifecycle satisfied by: feat__ai_platform
The platform shall manage AI model artifacts with unique version identification and shall support update and rollback of model artifacts independently of the application software.
feat_req__ai_platform__model_lifecycle		Feature Requirement

AI Workload Metrics		status: valid security: NO safety: QM
reqtype: Non-Functional valid_from: v2.0.0 version: 1 derived from: stkh_req__ai_platform__observability satisfied by: feat__ai_platform
The platform shall report inference latency, memory consumption and inference errors of AI workloads via the existing logging and tracing mechanisms.
feat_req__ai_platform__observability		Feature Requirement