Statements (50)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:machine_learning_infrastructure
|
| gptkbp:deployment |
2016
|
| gptkbp:developedBy |
gptkb:Uber
|
| gptkbp:documentation |
https://eng.uber.com/michelangelo-machine-learning-platform/
https://eng.uber.com/ludwig/ https://eng.uber.com/petastorm/ https://eng.uber.com/horovod-open-source-distributed-deep-learning/ |
| gptkbp:enables |
gptkb:feature_store
A/B testing model monitoring batch inference model versioning real-time inference automated model retraining scalable distributed training |
| gptkbp:includes |
gptkb:Michelangelo
gptkb:Horovod |
| gptkbp:integratesWith |
gptkb:TensorFlow
gptkb:Apache_Spark gptkb:Docker gptkb:Kubernetes gptkb:PyTorch |
| gptkbp:location |
gptkb:San_Francisco,_California
|
| gptkbp:openSourcedComponent |
gptkb:Ludwig
gptkb:Horovod Petastorm |
| gptkbp:purpose |
support machine learning workflows at Uber
|
| gptkbp:scalesTo |
millions of predictions per second
thousands of models |
| gptkbp:supports |
online learning
feature engineering hyperparameter tuning data labeling model deployment data validation model monitoring model validation model explainability model training offline learning |
| gptkbp:usedBy |
gptkb:Uber_Eats
gptkb:Uber_Freight gptkb:Uber_Rides |
| gptkbp:usedFor |
fraud detection
pricing optimization ETA prediction matching riders and drivers |
| gptkbp:bfsParent |
gptkb:Michelangelo_Feature_Store
|
| gptkbp:bfsLayer |
8
|
| https://www.w3.org/2000/01/rdf-schema#label |
Uber ML infrastructure
|