Vertex AI

What is Vertex AI?

Vertex AI is Google Cloud's unified platform for building, deploying and managing machine learning and generative AI applications. Google's page now presents it under the Gemini Enterprise Agent Platform name.

Key capabilities of Vertex AI

• Model Garden: catalog of Google foundation models and third-party and open models
• Gemini access: multimodal models for text, images, code and audio, with tuning options
• Agent Builder: low-code tools to create and orchestrate AI agents
• AutoML: trains custom models on tabular, image, text and video data
• MLOps tooling: experiment tracking, model registry, monitoring and pipelines
• Data controls: data residency options plus encryption in transit and at rest

How Vertex AI works

Teams pick a model from Model Garden or train their own with AutoML or custom code, then deploy it to managed endpoints. Generative models are called by API and billed per token, while training and serving are billed by compute instance. Agent Builder connects models to business systems and APIs for multi-step tasks.

Who uses Vertex AI?

Data scientists, ML engineers and enterprise application teams on Google Cloud use it. Google lists ISO and SOC certifications, HIPAA support, GDPR features and data residency controls for regulated industries.

Vertex AI pricing

Vertex AI is pay as you go. Generative models are priced per token and training and serving by compute instance and storage. Google offers $300 in free credits to eligible new Google Cloud customers, typically valid for 90 days.

Vertex AI alternatives

Alternatives include Hugging Face for open model hosting, ZenML for pipeline orchestration, Hopsworks for feature stores, Evidently AI for monitoring, and Zep for agent memory.