A domain-specific fine-tuned model needs managed hosting. Which service fits?
Why B is the answer
Amazon SageMaker AI
Amazon SageMaker AI provides managed hosting for fine-tuned, domain-specific models.
In detail
- SageMaker AI provides managed hosting for custom models.
- You choose the hosting configuration.
Use case
An illustrative scenario, not a real incident.
A medical team fine-tuned a model and hosts it on a SageMaker AI endpoint.
See it in different ways
flowchart TD
A{"What do you need?"} -->|"Variable traffic, managed models"| B["Bedrock on-demand"]
A -->|"Consistent high throughput"| C["Bedrock Provisioned Throughput"]
A -->|"Your own model container"| D["SageMaker AI endpoint"]
A -->|"Offline volume"| E["Batch inference"]mindmap
root((Inference))
On demand
Variable traffic
Provisioned Throughput
Steady and high
Batch
Offline and cheaper
Cross-Region
Spreads peaks
Streaming
Show text as it arrivesflowchart LR S0["Fine-tune"] --> S1["Deploy to SageMaker AI"] --> S2["Call the endpoint"]
Key point: SageMaker AI hosts custom models.
Why the other options are wrong
- A. Amazon Kendra
- Kendra is enterprise search.
- C. AWS Glue
- Glue is for ETL.
- D. Amazon EMR
- EMR runs big data frameworks.