slug: edge-impulse
provider: Edge Impulse
generated_by: planning/capability-mapping/scripts/classify_capabilities.py
model: claude-opus-5
frame:
- Software & Technology
min_confidence: 0.7
capability_model:
  source: https://github.com/vincentmakes/turbo-ea-capabilities
  license: CC-BY-4.0
  attribution: Turbo EA Capabilities by Vincent Verdet — Turbo EA, https://github.com/vincentmakes/turbo-ea-capabilities,
    CC BY 4.0
  notice: NOTICE
edge_count: 7
edges:
- tag: Learn
  spec_file: edge-impulse-learn-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.9
  evidence: GET /api/{projectId}/training/keras/{learnId} getKeras Keras information; POST /api/{projectId}/training/keras/{learnId}
    setKeras Keras settings; GET /api/{projectId}/training/anomaly/{learnId}/metadata getAnomalyMetadata
  reason: Model training configuration, Keras/anomaly model settings, training data and metadata retrieval
    — squarely ML model lifecycle / MLOps under Artificial Intelligence Management.
- tag: Classify
  spec_file: edge-impulse-classify-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.85
  evidence: POST /api/{projectId}/classify/v2/{sampleId} classifySampleV2 Classify sample; GET /api/{projectId}/classify/all/metrics
    getClassifyMetricsAllVariants Get metrics for all available model variants; schemas ModelPrediction,
    BoundingBoxWithScore
  reason: Operations run trained ML models against samples and return predictions/metrics per model variant
    — inference and model evaluation within the ML model lifecycle (BC-610.60 Artificial Intelligence
    Management).
- tag: Optimization
  spec_file: edge-impulse-optimization-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.82
  evidence: POST /api/{projectId}/optimize/{jobId}/create-trial  createTrial  Create trial; GET /api/{projectId}/optimize/space  getSpace  Search
    space; 'Retrieves the EON tuner state'
  reason: Hyperparameter/architecture search over learn blocks and DSP blocks with trials, scoring and
    search space — ML model lifecycle work, i.e. AI/ML management, not generic optimisation.
- tag: DSP
  spec_file: edge-impulse-dsp-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.8
  evidence: GET /api/{projectId}/dsp/{dspId}/features/importance getDspFeatureImportance Feature importance;
    GET .../features/labels Feature labels; POST .../raw-data/{sampleId}/slice/run Get processed sample
    (slice)
  reason: Configuration and execution of signal-processing/feature-extraction blocks in the ML pipeline
    — feature engineering, part of ML model lifecycle / MLOps under Artificial Intelligence Management.
- tag: Impulse
  spec_file: edge-impulse-impulse-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.8
  evidence: POST /api/{projectId}/impulse createImpulse Create impulse; GET /api/{projectId}/impulses-detailed
    getAllDetailedImpulses Get all impulses (incl. metrics); schemas ImpulseDspBlock, TransferLearningModel,
    KerasModelMetadataMetrics
  reason: An 'impulse' is Edge Impulse's ML pipeline definition (DSP + learn blocks). CRUD over pipeline
    configurations and their metrics is ML model lifecycle management.
- tag: PerformanceCalibration
  spec_file: edge-impulse-performancecalibration-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.78
  evidence: POST /api/{projectId}/performance-calibration/parameters  setPerformanceCalibrationSavedParameters  Save
    performance calibration parameters; schemas PerformanceCalibrationGroundTruth, PerformanceCalibrationFalsePositive
  reason: Tunes and stores model post-processing parameters against ground truth and false-positive measurements
    — evaluation and tuning within the ML model lifecycle.
- tag: Deployment
  spec_file: edge-impulse-deployment-api-openapi.yml
  capability_id: BC-610.60
  capability_id_l1: BC-610
  capability_name: Artificial Intelligence Management
  confidence: 0.7
  evidence: GET /api/deployment/targets listAllDeploymentTargets Deployment targets; GET /api/{projectId}/deployment/download
    downloadBuild Download; GetLastDeploymentBuildResponse
  reason: Builds and downloads a trained model artefact for a chosen hardware deployment target — model
    packaging/deployment stage of the ML lifecycle. Could also read as software release engineering, hence
    moderated confidence; the artefact deployed is an ML model, so AI Management is the better fit.