Title + requirements
{
"role": "Senior Applied AI Engineer",
"posting": "Build production agents and RAG workflows. Own Python services, React/TypeScript applications, and enterprise delivery on AWS."
}PUBLIC CAREER EVIDENCE · AGENT INTERFACE
Bring a role.
Get the evidence.
A browser agent can compare a job with my documented experience and receive a structured assessment, with sources and honest gaps.
01 / CONNECT
Use the portfolio’s
browser tool.
Open the portfolio in a browser that supports WebMCP. The page exposes assess_role_fit, including on the welcome screen. Complete Google sign-in in that browser before using AI.
The tool uses the same account credits and guardrails as the visible form. A page visit never starts an assessment.
- Discover.Read the manifest or inspect the page’s available browser tools.
- Provide the role.Submit a title and requirements, or a public HTTPS listing URL.
- Compare.The role goes directly to Decisions API scoring. An agent explains the supporting evidence.
- Read the result.The assessment is returned as JSON and shown in the portfolio’s role-match panel.
02 / REQUEST
Two ways to bring the details.
Public listing URL
{
"job_url": "https://company.example/careers/ai-engineer"
}URL import extracts the title, description, and evidenced salary when available, then runs the assessment. Both operations use AI credits. If a site blocks fetching, paste its requirements instead.
A title alone prepares the form and requests requirements, without using AI. Titles are limited to 160 characters; descriptions need 40–12,000 characters. Use a description or a URL, rather than both.
03 / RESPONSE
A score with something behind it.
Documented fit
A 0–100 evidence score, relevant delivery, capabilities, and depth.
Sources and gaps
Cited public evidence, supported strengths, and requirements worth a conversation.
Compensation context
Published salary, when available, remains separate from the evidence score. No acceptance probability is invented.
This is a portfolio comparison, not a hiring decision. AI can make mistakes; check the cited sources.
ACCESS & DATA
One public purpose.
The interface exposes public career matching. It has no access to private owner data, Gmail, visitor identities, or administrative tools. Submitted job text is not saved by this application; OpenAI processes it to generate the assessment. Usage metadata follows the privacy policy. Remove confidential information before submitting.
Unsigned requests return SIGN_IN_REQUIRED. Exhausted credits, invalid input, unavailable services, and cancellation return structured statuses. Honor retry guidance and avoid automatic retry loops.
WebMCP is a browser interface, available while this page’s portfolio is open in a compatible browser. There is no anonymous remote MCP or A2A server. Other browsers can use the role-match form after Google sign-in. Keep session cookies in the browser; no OpenAI key is needed.