Model Context Protocol
Ask your AI about me.
This portfolio is also an MCP server. Connect your AI assistant and it can read my profile, experience, projects and blog directly, then answer questions or assess fit for a role.
Works with any MCP client
Connect
Endpoint: https://your-domain/mcp (Streamable HTTP · read-only · no auth)
Claude Code
claude mcp add --transport http madhukar https://your-domain/mcp
Claude (web & desktop)
Settings → Connectors → Add custom connector URL: https://your-domain/mcp
Cursor · .cursor/mcp.json
{
"mcpServers": {
"madhukar": {
"url": "https://your-domain/mcp"
}
}
}VS Code · .vscode/mcp.json
{
"servers": {
"madhukar": {
"type": "http",
"url": "https://your-domain/mcp"
}
}
}7 tools · resources · 1 prompt
What your AI can do
Tools
get_profileSummary, contact details and availabilityget_experienceCurrent role, highlights and key contributionsget_skillsTech stack by category, or check a single technologylist_projectsKey projects and Lab experiments, filterablesearch_blogRanked keyword search over technical postslist_blog_postsEvery post, newest firstget_blog_postFull text of a post as Markdown
Prompt
assess_fitPaste a job description; your AI gathers evidence with the tools and reports matches and gaps.
Resources
portfolio://profile and portfolio://blog/{slug}
Try asking
- “Has Madhukar built production RAG systems? Cite projects.”
- “Use the assess_fit prompt with this job description: …”
- “Summarise his blog post about hybrid search in 5 bullets.”
- “Does he know Redis and Kubernetes?”
Calls search_blog on this server