aws-samples / aws-samples/aws-generativeai-partner-samples
Proposal: add a databricks/ partner section — Genie as a governed MCP tool for Amazon Bedrock AgentCore (Energy & Utilities)
- Dominant language
- Jupyter Notebook
- Stars
- 73
- Forks
- 50
- PR merge metrics
- No merged PRs in 30d
Description
Hi maintainers 👋
Following CONTRIBUTING ("open an issue to discuss significant work"), I'd like to propose adding a new **`databricks/`** partner section. Databricks is an AWS partner and there isn't a Databricks folder yet (Snowflake, MongoDB, Confluent, Elastic, etc. are represented), so I'd like to align on placement/format before opening a PR.
**Proposed first sample — Databricks Genie as a governed MCP tool for Amazon Bedrock AgentCore (Energy & Utilities use case):**
- An Amazon Bedrock AgentCore agent calls **Databricks Genie** over **MCP** to get governed, natural-language → SQL answers over a lakehouse — governed by Unity Catalog, with no data movement.
- Use case: **Energy & Utilities** smart-meter load forecasting on **synthetic** data — a vertical not currently represented in the repo.
- Structure would mirror `snowflake/mcp/`: `mcp-servers/`, `mcp-clients/`, `data/` (synthetic), a README, and an architecture diagram; notebook-friendly per the repo convention.
**Scoping questions:**
1. Preferred layout/naming for a new partner section — `databricks/mcp/...` like Snowflake, or a different convention?
2. Notebook-first vs. scripts + README — any preference for new samples?
3. Any constraints to know up front (dependencies, region/service assumptions, test expectations)?
I'll follow CONTRIBUTING (fork, focused PR, MIT-0, CLA) once we've aligned on the shape. Thanks!
Contributor guide
Research direction
Start with CONTRIBUTING and the existing snowflake/mcp/ section to compare partner layout and sample conventions. Resolve the proposed databricks/ structure, notebook-versus-script format, dependencies, region assumptions, and test expectations before implementing; done means an agreed focused PR containing the described synthetic Energy & Utilities sample.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- aws
- Domain
- ai, cloud
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100