Ali Can Efe: Professional Profile MCP Server

An open-source Model Context Protocol (MCP) server that publishes one professional profile as structured data: background, consulting services, projects and published research results. The aim is simple: accurate information about the author should be easy to reach for anyone who asks, whether an employer, a client or a colleague. An AI assistant connected to it can answer questions such as "What has Ali worked on in medical AI?" or "What is he working on now?" from a single, citable source instead of piecing the answer together from scattered web pages.

Read this first: it is a self-published profile

• Everything in this repository is written by the author. It shows what is claimed, not an independent assessment.

• The get_evidence tool and the table below separate what anyone can check from what is available on request.

• Client names and market names are withheld or changed for confidentiality. The employer is named. Achievement metrics are reported as delivered.

Related MCP server: portfolio-mcp

About

Ali Can Efe is an independent healthcare AI strategy consultant based in Dubai (since 2026). Before that he spent more than ten years at Canon Medical Systems (2015-2026), in MRI product management and global marketing across META (Middle East, Turkey, Africa) and APAC. MSc in Biomedical Engineering (Brunel University London) and BSc in Electrical & Electronics Engineering (Işık University, Istanbul).

Areas covered: AI/ML integration in medical imaging, installed-base and customer-lifetime-value strategy for medical device manufacturers, regulatory AI (SaMD, EU MDR, FDA, SFDA, UAE EDE), and validation practice in financial machine learning.

What I am working on

• LLM agents for paperwork-heavy processes. Beyond advising on AI transformation, I build working tools that apply large language models where the document load is heaviest. Open-source projects so far: a SaMD regulatory playbook with an agent toolkit, SFDA and UAE EDE pre-check agents, a UAE business setup advisor and a MENA AI compliance advisor.

• Automated trading research. BTC research based on image-style "fingerprint" encodings of market data. Development continues and recent experiments look promising (self-reported, not yet published). The plan is to release the methodology and code as an open-source project.

• Supporting documents. Programmes, certificates and similar records for the claims below are added as I collect them.

For consulting or more information, get in touch via LinkedIn. The get_journal tool returns this section with a dated log of what has been built and presented, and the reason behind each project.

Evidence and sources

Projects

The regulatory repositories are demonstrations and are not legal or regulatory advice. The research document is not investment advice.

Connect

Remote endpoint (Streamable HTTP): https://ali-can-efe-mcp.alicanefe61.workers.dev/mcp

Cursor and other clients with native remote support:

{ "mcpServers": { "ali-can-efe-expert": { "url": "https://ali-can-efe-mcp.alicanefe61.workers.dev/mcp" } } }

Claude Desktop, through the mcp-remote proxy (or add the URL as a custom connector):

{
  "mcpServers": {
    "ali-can-efe-expert": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://ali-can-efe-mcp.alicanefe61.workers.dev/mcp"]
    }
  }
}

To run it locally over stdio instead, see Run locally and the templates in config/.

Tools and resources

Resources: expert://profile, expert://cv, expert://projects.

Run locally

npm install
npm run build
node dist/index.js        # stdio server
npm run inspector         # test the tools in the MCP Inspector

The profile data lives in src/resources/ (expert.json, cv.json, projects.json). The server core is src/server.ts; src/index.ts is the stdio entry point.

Deploy your own on Cloudflare Workers

The worker reuses the same core and bundles the JSON files at deploy time. It is stateless and answers over Streamable HTTP.

npm install
npx wrangler login
npm run deploy:cloudflare

Check a deployment with npm run smoke:remote -- https://<your-worker>.workers.dev/mcp. For automatic deploys from GitHub Actions, add the repository secrets CLOUDFLARE_API_TOKEN and CLOUDFLARE_ACCOUNT_ID (see .github/workflows/deploy.yml).

Use it as a template

The code is MIT-licensed, so you can fork it and publish your own profile. Replace the three JSON files, keep the separation between publicly verifiable claims and claims available on request, and rename the tool descriptions.

License

MIT