GCP.
Google Cloud Platform is a cloud provider strongest in data and machine learning workloads — BigQuery, Dataflow and Vertex AI — and common in organisations whose analytics or AI work pulled them there before their applications followed.
Where GCP fits on a long engagement.
GCP is frequently adopted sideways. A team starts with BigQuery for analytics, then adds storage, then a service, and eventually there is a production estate nobody planned. Project structure and IAM decided retrospectively are harder to fix than they would have been to set.
Its data services are genuinely strong, and that is usually the reason to be there. Where an organisation is on GCP for analytics but runs its applications elsewhere, the sensible question is whether the split is deliberate or just how it happened.
What an assigned team does with GCP.
GCP estates frequently start in analytics and grow into production without anyone deciding they should. Project structure and IAM set retrospectively are harder to correct than to establish, and the correction is the sort of work that never wins against a feature unless it is separately assigned.
Where the estate genuinely exists to serve models rather than applications, the binding constraint is usually operational, and capacity is better assigned under MLOps engineering than under general platform work.
What we use GCP for.
- Project and IAM structure Boundaries set deliberately so environments are separated and permissions can be reasoned about.
- Data platform operations BigQuery, pipelines and scheduling run as a platform rather than as a collection of one-off jobs.
- Cost attribution Labels and budgets applied so analytical spend can be traced to the team generating it.
How GCP capacity is assigned.
GCP work is assigned under DevOps managed services, and where the estate exists mainly to serve analytics, alongside data capacity on one agreement.
Tell us what your roadmap needs GCP for.
A service delivery manager replies with the disciplines we would assign, the monthly capacity and what the first month looks like.
Loading the contact form… You can also email hello@azendo.co.
We reply within one working day. No obligation, and no newsletter.