Skip to main content

GCP Ai Platform Custom Job

A Vertex AI / AI Platform Custom Job represents an ad-hoc machine-learning workload that you want Google Cloud to run on managed infrastructure. By pointing the job at a custom container image or a Python package, you can execute training, hyper-parameter tuning or batch-processing logic with fine-grained control over machine types, accelerators, networking and encryption. The job definition is submitted to the projects.locations.customJobs API and Google Cloud provisions the required compute, streams logs, stores artefacts and tears the resources down once the job finishes.
Official documentation: https://cloud.google.com/vertex-ai/docs/training/create-custom-job

Supported Methods​

  • GET: Get a gcp-ai-platform-custom-job by its "name"
  • LIST: List all gcp-ai-platform-custom-job
  • SEARCH

gcp-ai-platform-model​

A successful Custom Job can optionally call model.upload() or configure model_to_upload, causing Vertex AI to register a Model resource containing the trained artefacts. Overmind links the job to the resulting gcp-ai-platform-model so you can trace how the model was produced.

gcp-artifact-registry-docker-image​

Custom Jobs usually run inside user-supplied container images. When the image is stored in Artifact Registry, Overmind records a link between the job and the specific gcp-artifact-registry-docker-image it pulled, making it easy to audit code and dependency provenance.

gcp-cloud-kms-crypto-key​

If you enable customer-managed encryption keys (CMEK) for the job, Google Cloud encrypts logs, checkpoints and model files with the specified KMS key. The job therefore references a gcp-cloud-kms-crypto-key, which Overmind surfaces to highlight encryption dependencies and key-rotation risks.

gcp-compute-network​

Custom Jobs can be configured to run on a private VPC network (VPC-SC or VPC-hosted training). In that case the job is associated with the chosen gcp-compute-network, allowing Overmind to show ingress/egress paths and potential network exposure.

gcp-iam-service-account​

Vertex AI executes the workload under a user-specified or default service account. The job’s permissions—and hence its ability to read data, write artefacts or call other Google APIs—are determined by this gcp-iam-service-account. Overmind links them to flag overly-privileged identities.

gcp-storage-bucket​

Training data, intermediate checkpoints and exported models are commonly read from or written to Cloud Storage. The Custom Job specifies bucket URIs (e.g., gs://my-dataset/*, gs://my-model-output/). Overmind connects the job to each referenced gcp-storage-bucket so you can assess data residency and access controls.