Add AI Jobs Signal endpoint: ADS scorer, MCP detector, taxonomy and baseline

[Trailgenic]

Motivation

• Provide a serverless API to score labor-market job postings for AI Deployment Signals (ADS) and detect MCP signals to track market adoption and tier shifts.
• Standardize tier taxonomy and scoring logic so downstream tooling can compare current feed samples against a stored market baseline.
• Enable quadrant positioning by combining ADS with ARI for quick leader/laggard/inflation heuristics.

Description

• Added a new API handler api/ai-jobs-signal.js that fetches job postings from Indeed RSS, extracts company and skills, classifies postings, computes ADS metrics, and returns a structured payload including scoring, tier_distribution, top_postings, benchmark_delta, and positioning via assignQuadrant.
• Implemented classifier and scoring helpers in shared/ads-classifier.js which call the Anthropic model to classify postings with a JSON-only system prompt and provide computeADS and assignQuadrant helpers; fallback heuristics are applied when the model is unavailable.
• Added taxonomy and signal lists in shared/ads-taxonomy.js with TIER_KEYWORDS, signal keyword arrays, and getTierForSkill helper for deterministic tier lookups.
• Added a market shared/ads-baseline.json containing a March 2026 benchmark used for benchmark and compare modes and delta calculations.
• Updated vercel.json routing to expose the new /api/ai-jobs-signal endpoint.

Testing

• No automated tests were added or executed as part of this change; CI/unit tests were not modified by this PR.

---

Codex Task

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
exmxc-audit	Ready	Preview, Comment	Mar 30, 2026 5:03pm