Practice free →
HomeGCP ML Engineercloudcomputingmletraining › Which compute option should an ML Engineer choos…

Which compute option should an ML Engineer choose for an INFERENCE workload on a low-power edge device (e.g. mobile, IoT) — referenced in PMLE §3.3 + §4.2?

ACloud Memorystore
BEdge TPU / Coral devices (or quantised TensorFlow Lite / ONNX models on the device)
CCloud DNS
DVertex AI Prediction running in a Google datacenter only
Answer & Solution
Correct answer: B. Edge TPU / Coral devices (or quantised TensorFlow Lite / ONNX models on the device)
Edge TPU + TF Lite is the edge-inference path (per PMLE §3.3, §4.2). The other options aren't edge.
Solve this in the app — GCP ML Engineer practice & 24k+ MCQs →
Related questions