The Problem Nobody Talks About
Every week someone in the AI community asks: "Do you know a good agentic engineer?" And every week the answer is the same: "Yes, but they're not on LinkedIn."
The people exist. The infrastructure does not.
Traditional job boards were designed for a world of PDF resumes and 2-week hiring cycles. That world is ending. 70% of employers have already moved to skills-based hiring. 53% dropped degree requirements entirely. 99% use AI somewhere in their hiring process.
Yet the tools have not changed.
Why This Matters Now
March 2026 data points:
- 97M+ monthly MCP SDK downloads
- 5,800+ MCP servers in production
- Anthropic, OpenAI, Google, Microsoft — all shipping MCP clients
- Agentic Engineer salaries hitting $200K+
The Model Context Protocol is becoming the universal interface for AI tooling. Every serious AI application ships MCP support. Datadog went GA with their MCP server this month. The tooling layer is mature.
We're past "early adopter" territory. This is infrastructure now.
What MCP-Native Hiring Could Look Like
Here's the thought experiment we started with:
What if your MCP server was your resume?
Not a list of past jobs. Not a skills keyword cloud. But actual, running code — tools, integrations, context you've built — accessible to any AI agent that wants to evaluate you.
Employers describe what they need in natural language. AI agents search candidates based on what they actually build. Skills-based matching, not keyword matching.
That's the core idea behind MCPHire: a job marketplace where MCP is a first-class citizen, not an afterthought.
The Market Signal
AI recruiting is a $661M market today, on its way to $1.12B by 2030. Agentic roles are the fastest-growing segment. Yet the primary tools for finding these people are: LinkedIn (built for 2003), Greenhouse (built for 2012), and warm intros.
The people building the future are being hired like it's still the past.
Where We Are
Honest answer: early. Very early.
We have a demo live at mcphire.com. We're working through the architecture for an actual MCP-native matching layer. No promises on launch dates.
What we do know: the problem is real. The 200K+ agentic engineers who get hired through warm intros because no job board speaks their language. The companies scaling AI teams who spend months searching for someone who already exists in their extended network.
If this problem resonates — follow along. Building in public.
MCPHire — job marketplace for the agentic AI era. Pre-beta, building in public.
Top comments (0)
Some comments may only be visible to logged-in visitors. Sign in to view all comments.