[Feature]: Batch API endpoint for creating judgements

### Problem Statement

When creating multiple judgements, especially using an LLM and MCP, sending individual create judgement requests becomes inefficient.

### Proposed Solution

This feature would add a batch create to the [judgements API](https://elastic.github.io/relevance-studio/#/docs/latest/reference/rest-api?id=judgements-api).

One solution would be to allow an alternate payload format in the existing endpoint, e.g. :

```
{
"judgements": [
{
"scenario_id": "67d554e2-8622-4c6c-9601-bf0e0a06128d",
"index": "demo",
"doc_id": "123",
"rating": 3
}
]
}
```

### Alternatives Considered

_No response_

### Enhancement Area

Frontend (UI)

### Priority

Critical - Blocking my workflow

### Use Case

_No response_

### Additional Context

_No response_

Start with the judgements API documentation and the existing create-judgement endpoint. Trace its request validation and tests, then determine how the proposed judgements array should be accepted and verified. Done means one request creates all supplied judgements with the documented payload and coverage for the batch behavior.

Written by the indexing model from the issue text.