$ fab auth login
$ fab ls
workspaces/ capacities/
$ fab job start team.workspace/pipeline.datapipeline- Role
- Senior Product Manager
- Dates
- 2024-Present
- Status
- Current work
- Context
- Work at Microsoft
Microsoft Fabric developer tools
Public product work on Fabric CLI, Fabric MCP, developer automation, and interfaces for AI-assisted development.
The case in five moves
Two public developer tools, one underlying product question: how should people and AI agents build on a complex platform without carrying its full complexity?
- Constraint
- Fabric exposes a broad surface of public APIs and item definitions. Developers needed a faster, more predictable way to explore, automate, and generate correct integration code.
- Leverage point
- Use familiar interfaces as shared rails. A file-system-inspired CLI supports people and automation; an MCP server supplies agents with official API context and schemas.
- What was built
- Fabric CLI for interactive and scripted workflows, and Fabric MCP for local, read-only access to API specifications, item schemas, and recommended patterns.
- Evidence
- Fabric CLI moved from preview to general availability and then open source. Fabric MCP launched publicly as an open-source preview within Microsoft’s MCP repository.
- What I learned
- The interface is not a thin layer over complexity. A coherent interface becomes platform infrastructure when developers, automation, and agents can all rely on the same model.
Build notes
Since August 2024, I have worked on Microsoft Fabric’s Core Platform team, now as a Senior Product Manager, focused on developer automation and AI-assisted development.
Fabric CLI
I lead product work on Fabric CLI, a file-system-inspired command-line interface for exploring and automating Microsoft Fabric. The product moved from preview to general availability and then open source.
Public product writing:
Fabric MCP
I also lead product work across Fabric MCP, which gives AI agents structured context about Fabric’s public APIs, item definitions, and recommended patterns. It is open source and designed to keep execution under the developer’s control.
What connects the work
The common problem is bigger than one interface. Developers and AI agents need predictable ways to understand, automate, and build on a complex platform. My role is to turn that need into coherent products, shared patterns, and clear contribution paths.
This page describes public professional work. The views and independent projects on this website are personal and do not represent Microsoft.