feat: AI wrapper for AWS Bedrock and GCP Garden integration #3

Overview

Create an abstraction layer (AI wrapper) that enables teams to easily utilize large language models and other AI services via AWS Bedrock or GCP Garden, allowing for flexible backend selection and seamless integration into our existing workflows.

Motivation

• Multi-cloud flexibility: Teams may have preferences or requirements to use either AWS Bedrock or GCP Garden for AI/ML workloads.
• Unified interface: A common API/wrapper will reduce code duplication and make it easier to switch between providers.
• Future extensibility: Lays the groundwork for supporting additional AI providers in the future.

Requirements

• Design a common interface for AI model invocation (e.g., text generation, embeddings, etc.).
• Implement provider-specific adapters for AWS Bedrock and GCP Garden.
• Allow configuration to select the desired backend at runtime or via environment/config files.
• Provide clear error handling and logging for unsupported features or provider-specific issues.
• Include documentation and usage examples for both providers.

Acceptance Criteria

• Teams can switch between AWS Bedrock and GCP Garden with minimal code changes.
• The wrapper supports at least basic text generation for both providers.
• Configuration is straightforward and well-documented.
• Unit/integration tests cover both backends.

Additional Context

• AWS Bedrock documentation
• GCP Garden documentation