Mike is Hiring

Source-tracked crypto and frontier-AI jobs, explainable hiring intelligence, and candidate tools across job, company, market, and career surfaces.

Created by Mike Williams, Flashbots' founding recruiter and a builder of evidence-led hiring systems for frontier technical teams. Public identity: LinkedIn · GitHub · X.

🤖 Hosted MCP server (no install)

This repo hosts a production-ready Model Context Protocol server for Web3 hiring intelligence — 60 tools covering job search, candidate fit ranking, salary bands, company comparison, market signals, AI hiring evidence provenance, private candidate workspaces, and agent handoffs.

Run the local stdio server in one step:

npx -y mikeishiring-mcp

Or connect an MCP client directly to the hosted endpoint with no package install.

claude mcp add --transport http web3-jobs https://www.mikeishiring.com/api/public/v1/mcp

• Catalog + live demo + connect instructions: https://www.mikeishiring.com/mcp
• JSON-RPC endpoint: POST https://www.mikeishiring.com/api/public/v1/mcp
• Health probe: GET https://www.mikeishiring.com/api/public/v1/mcp/health
• 45-assertion contract test runs on every push (scripts/test-mcp-http.mjs)
• Architecture map: docs/ARCHITECTURE.md
• Hosted entry: frontend/v2/functions/api/public/v1/mcp.ts (Cloudflare Pages Function)
• Local stdio: npx -y mikeishiring-mcp (Node 22.18+)
• Source: mcp/web3-jobs-public-api.mjs

The MCP is honest about data quality: results that hit the skill-bloat penalty surface a pool_quality.warning; queries that don't resolve to a canonical role surface a query_resolution_warning; salary band numbers come with display_short ($175k–220k) for one-shot rendering. See docs/ARCHITECTURE.md "Open items" for the upstream data ceilings.

📖 Key Documentation

Start here before diving into code:

| Document | Purpose | When to Read | |----------|---------|--------------| | docs/PRODUCT_CONTRACT.md | Canonical source of truth - Mission, personas, Evidence Ladder, canonical objects, phase plan | Before any major change | | docs/strategy/AI_TALENT_INTELLIGENCE_ROADMAP.md | Active roadmap for making Frontier AI talent/hiring intelligence first-class | Planning AI talent work | | docs/talent-intelligence/README.md | Current public entrypoint for AI talent readers, trust boundaries, artifact navigation, and refresh commands | Starting the AI talent intelligence pack | | docs/talent-intelligence/AI_HIRING_INTELLIGENCE_METHODS.md | Methodology, evidence ladder, promotion rules, privacy boundaries, and reproducible refresh commands for AI hiring intelligence | Auditing how AI hiring claims are produced | | docs/talent-intelligence/AI_TALENT_REPUTATION_SCORECARD.md | Reputation-readiness scorecard for serious AI engineers and researchers | Auditing AI talent credibility | | docs/talent-intelligence/AI_ROLE_READINESS_MATRIX_2026-07-07.md | Generated role-family readiness matrix translating AI candidate role bars into proof artifacts, weak signals, prep actions, and work-vector questions | Auditing what current source-backed AI roles appear to evaluate | | docs/talent-intelligence/AI_RESEARCH_ENGINEERING_INTERVIEW_PREP_KIT_2026-07-07.md | Generated candidate-facing prep kit translating AI candidate role bars into portfolio artifacts, work-vector questions, weak signals, and source trails | Preparing for a source-backed AI candidate path | | docs/talent-intelligence/AI_HIRING_MARKET_REPORT_2026-07.md | Monthly AI hiring market report with lab comparison matrix, interview pattern taxonomy, role-family pressure, evaluation signals, and source/privacy policy | Reading the public AI hiring market readout | | docs/talent-intelligence/AI_HIRING_EVIDENCE_PROMOTION_QUEUE_2026-07-07.md | Generated promotion-readiness queue for lower-tier media, forum, interview-prep, LinkedIn, and X/social evidence | Auditing which public-web rows can be shown, which stay research-only, and why | | docs/talent-intelligence/AI_HIRING_EVIDENCE_LEDGER_2026-07-07.md + JSON + schema | Generated and queryable source provenance ledger across official lab rows, research bars, work vectors, external evidence, and social research leads | Auditing the source trail behind public AI hiring claims | | docs/talent-intelligence/OPENAI_ANTHROPIC_HIRING_DEEP_DIVE_2026-07-07.md | Generated source-tiered OpenAI vs Anthropic hiring deep dive with stage maps, assessment/AI-use boundaries, proof portfolios, role bars, and lower-tier online evidence handling | Reading the focused OpenAI/Anthropic hiring research brief | | docs/talent-intelligence/AI_TALENT_PUBLIC_READINESS_AUDIT_2026-07-07.md | Generated acceptance matrix for what the public AI talent pack proves, what remains conditional, and what external reputation still requires | Auditing whether the pack is ready for serious AI readers | | docs/talent-intelligence/AI_TALENT_EXTERNAL_REVIEW_BRIEF_2026-07-07.md | Generated skeptical-reader brief for what the public AI talent pack proves, what stays limited, and which commands reproduce the evidence | Reviewing the pack like an external AI engineer or researcher | | docs/talent-intelligence/AI_TALENT_EXTERNAL_REVIEW_LEDGER_2026-07-09.md + JSON | Generated external review ledger for reviewer targets, pending requests, accepted evidence rules, current zero accepted external reviews, intake fields, and reputation limits | Auditing the path from repository evidence to real outside reputation | | data/ai_talent_external_review_intake_contract.yaml + fixture | Validated intake contract for real AI engineer/researcher review feedback, consent, evidence type, public-safe projection, and no-overclaim gates | Turning external critique into repeatable provenance instead of one-off praise | | docs/talent-intelligence/AI_TALENT_ARTIFACT_MANIFEST_2026-07-07.md | Deterministic manifest with public artifact paths, generation commands, verification commands, file sizes, line counts, and SHA-256 hashes | Auditing the public AI talent pack as a reproducible release | | docs/talent-intelligence/AI_TALENT_SOURCE_REFRESH_PLAN_2026-07-08.md + JSON | Generated and queryable source-refresh plan for access methods, owners, cadence, stale clocks, promotion gates, component coverage, and thread refresh coverage | Auditing that sources are repeatable and gated rather than one-off scrapes | | docs/talent-intelligence/AI_TALENT_REFRESH_RUN_LEDGER_2026-07-08.md + JSON | Generated refresh-run ledger for source-row resolution, family decisions, held social leads, thread materialization, and blockers | Auditing the latest public-safe repeatability run | | docs/talent-intelligence/AI_TALENT_OPERATING_MODEL_2026-07-08.md + JSON | Generated and queryable operating model for source refresh, private employee/profile resolution, social-post classification, signal-thread materialization, public projection, and current blockers | Auditing how the AI talent graph runs as a repeatable system instead of a one-off scrape | | docs/talent-intelligence/AI_TALENT_SIGNAL_COMPONENT_MAP_2026-07-08.md + JSON | Generated and queryable component map for employees, affiliations, social profiles/posts, post types, thread families, provenance, repeatability, and privacy boundaries | Auditing how the AI talent graph is built without exposing raw person/profile/post rows | | docs/talent-intelligence/AI_TALENT_SIGNAL_THREAD_MAP_2026-07-08.md + JSON | Generated and queryable signal-thread map joining official hiring rows, named-source context, social leads, post types, provenance, caveats, and promotion blockers | Auditing how employee/social/profile modeling becomes public-safe hiring signal | | docs/talent-intelligence/AI_TALENT_SIGNAL_PROVENANCE_GRAPH_2026-07-08.md + JSON | Generated and queryable provenance graph connecting source rows, source IDs, source hosts, components, post types, refresh families, signal-thread families, thread instances, companies, and promotion blockers | Auditing the source-to-signal path with clear provenance and public-safe graph edges | | docs/talent-intelligence/FRONTIER_AI_HIRING_MAP_2026-07-07.md | Generated public Frontier AI hiring map from source-backed lab data | Reading or refreshing the public AI hiring market report | | docs/talent-intelligence/AI_TALENT_COVERAGE_GAP_REPORT_2026-07-07.md | Generated AI talent coverage report tracking 50+ AI companies, 10+ Frontier labs, and deep-pack gaps | Auditing AI company and lab breadth | | docs/strategy/FRONTIER_AI_HIRING_INTELLIGENCE_CONTRACT.md | Cross-domain contract for Frontier AI company, hiring, talent, and prediction packs | Before changing /frontier or AI data | | docs/VISION.md | Long-term vision and aspirational goals | Understanding mission & direction | | docs/MASTER_PLAN.md | Execution-oriented implementation plan | Planning work, checking status | | docs/ops/HIRING_HISTORY_RUNBOOK.md | History/status backfill and readiness runbook | When validating the hiring-history database | | docs/architecture/PROJECT_SURFACE_AUDIT.md | Active vs compatibility vs transitional repo surfaces | Before extending older paths | | docs/PROJECT_COMPONENT_CHECKLIST.md | Repo-wide component inventory and checklist | When mapping the system | | docs/guides/QUICKSTART.md | Fast onboarding (5 minutes) | Getting started quickly | | docs/guides/WORKFLOWS.md | Common workflows and commands | Day-to-day development | | docs/SECURITY_QUICK_REFERENCE.md | API security status - What's safe, what's not, quick fixes | Before public deployment |

Frontier AI seed data: data/frontier_lab_hiring_intelligence.yaml captures source-backed hiring-process, candidate-playbook, proof-portfolio, role-bar, and compensation/location evidence for first-tier AI labs.

External AI hiring evidence: data/frontier_lab_external_hiring_evidence.yaml captures social, media, interview-prep, and candidate-report signals with tiering and public-use guards.

AI candidate role bars: data/ai_research_engineering_bars.yaml captures source-backed global role families across technical, recruiting, sales, and design paths.

AI research work vectors: data/ai_research_work_vectors.yaml connects official lab research areas to role families, technical substrate, public artifacts, and evaluation questions.

Shared AI/Web3 page-depth contract: data/domain_intelligence_page_layers.yaml defines the required company-page layers for AI and Web3/crypto intelligence.

Web3 hiring playbooks: data/domain_hiring_playbooks.yaml defines role-lane demand signals, proof artifacts, story prompts, and prep actions for crypto candidates.

Web3 company hiring evidence: data/web3_company_hiring_evidence.yaml captures source-backed company-level hiring process, assessment, and candidate-proof evidence with public/private guards.

AI talent reputation gate:

npm run verify:ai-talent-reputation

Frontier AI hiring map report:

npm run export:frontier-ai-hiring-map
npm run export:ai-role-readiness-matrix
npm run export:ai-research-engineering-interview-prep-kit
npm run export:ai-hiring-market-report
npm run export:ai-hiring-evidence-promotion-queue
npm run export:ai-hiring-evidence-ledger
npm run export:ai-talent-public-readiness-audit
npm run export:openai-anthropic-hiring-deep-dive
npm run export:ai-talent-external-review-brief
npm run export:ai-talent-external-review-ledger
npm run export:ai-talent-source-refresh-plan
npm run export:ai-talent-authorized-capture-run
npm run export:ai-talent-refresh-run-ledger
npm run export:ai-talent-operating-model
npm run export:ai-talent-signal-component-map
npm run export:ai-talent-signal-thread-map
npm run export:ai-talent-signal-provenance-graph
npm run export:ai-talent-artifact-manifest
npm run query:ai-hiring-evidence-ledger -- --entity OpenAI --tier A --query interview
npm run query:ai-talent-signal-threads -- --company OpenAI

The generated AI_TALENT_AUTHORIZED_CAPTURE_RUN_2026-07-08.md and AI_TALENT_AUTHORIZED_CAPTURE_RUN_2026-07-08.json artifacts prove the authorized-capture lane as public-safe hashed post/source metadata. The machine-readable intake contract lives at data/ai_talent_authorized_capture_input_contract.yaml; validate a real authorized export before generation with python scripts/run_ai_talent_authorized_capture.py --validate-input-only --input <authorized-export.yaml>. The hosted and local MCP surfaces also expose get_ai_talent_source_refresh_plan for repeatable source-refresh metadata, get_ai_talent_authorized_capture_run for public-safe authorized social/profile/post capture runs, get_ai_talent_refresh_run_ledger for the current source-row/family/thread run state, get_ai_talent_operating_model for the end-to-end employee/profile/social-post operating model, get_ai_talent_signal_components for the component/repeatability map, get_ai_talent_external_review_ledger for reviewer targets, pending review requests, accepted evidence rules, and the no-overclaim reputation boundary, and get_ai_talent_signal_threads for bounded public-safe queries over woven signal threads, plus get_ai_talent_signal_provenance_graph for source-to-component, post-type, refresh-family, and signal-thread provenance edges. Use npm run validate:ai-talent-external-review-intake -- --input <external-review.yaml> to validate real reviewer feedback against data/ai_talent_external_review_intake_contract.yaml before it can be projected into the public external-review ledger. The public /api/v2/frontier/external-review-ledger route exposes the public-safe review ledger for target, reviewer-role, request-status, evidence-status, query, and limit review. The public /api/v2/frontier/talent-operating-model route exposes the generated operating model for stage, component, private-boundary, gap, and projection review. The public /api/v2/frontier/authorized-capture-run route exposes the public-safe authorized-capture projection for hashed post/source metadata, post-type classification, gate, and privacy review. The public /api/v2/frontier/refresh-run-ledger route exposes the latest public-safe refresh-run ledger for source-row, family, thread, blocker, and gate-decision review.

AI talent coverage gap report:

npm run export:ai-talent-coverage-gap-report
npm run verify:ai-talent-coverage

Local setup check:

npm run verify:local-setup

Market intelligence platform for the web3 ecosystem. Automated scraping, AI-powered enrichment, and a React dashboard tracking jobs, companies, and talent signals.

graph LR
    subgraph Pipeline["Scraping Pipeline"]
        YAML[companies.yaml<br/>50+ configs] --> PW[Playwright scrapers]
        PW --> ENR[Claude AI enrichment<br/>taxonomy + scoring]
    end

    subgraph Storage
        ENR --> DB[(SQLite + Supabase)]
    end

    subgraph API["Flask API :5000"]
        DB --> R[40+ endpoints<br/>jobs / companies / candidates<br/>analytics / market / alerts]
    end

    subgraph Frontend["React Dashboard :3000 / AI Hiring :3001"]
        R --> P1[Jobs + Search]
        R --> P2[Company Drilldown]
        R --> P3[Market Intelligence]
        R --> P4[Talent Analytics]
        R --> P5[Career Intelligence]
        R --> P6[Role Taxonomy]
    end

Quick Start

# Backend (reference implementation only — not what's deployed in production;
# production is Cloudflare Pages Functions + Supabase Edge, see docs/ARCHITECTURE.md)
pip install -r requirements.txt
python -m server.app   # Flask reference API on :5000, run from repo root

# Frontend
cd frontend/v2 && npm install && npm run dev   # Vite prints the active local URL

# Stable local AI Hiring / Frontier surface
npm run dev:frontier                         # http://127.0.0.1:3001/frontier
npm run verify:frontier-page:local           # run in another terminal once dev is up

If http://127.0.0.1:5173/frontier shows an "Ops Center" or "Supabase access required" screen, that port belongs to a different local project. Use the dev:frontier URL above, or pass a custom URL with:

npm run verify:frontier-page:local -- --url http://127.0.0.1:<port>/frontier

Stack

| Layer | Technology | |-------|-----------| | Scraping | Python 3.11+, Playwright, YAML configs | | Enrichment | Claude AI, Pydantic, custom taxonomy | | Backend | Cloudflare Pages Functions + Supabase Edge (production); Flask 3 + SQLite (Python reference implementation in reference/python-backend/, not deployed) | | Frontend | React 18, TypeScript (strict), Vite 5, Tailwind 3.4, React Query 5, Framer Motion, Recharts | | Database | SQLite (local) + Supabase (production) |

API Surface

The legacy /api/* v1 surface (/api/alerts, /api/admin/*, etc.) has been removed. Everything is now served under /api/v2/* (and /api/public/v1/* for the hosted MCP server). The full generated route inventory lives in docs/api/V2_ROUTE_INVENTORY.md, kept in sync by npm run verify:v2-routes.

| Blueprint | Key Endpoints | |-----------|--------------| | Jobs | /api/v2/jobs, /api/v2/jobs/:id | | Companies | /api/v2/companies, /api/v2/companies/:id/drill-down | | Talent | /api/v2/talent/search, /api/v2/talent/flow, /api/v2/talent-explorer/* | | Analytics | /api/v2/analytics/dashboard, /api/v2/analytics/skill-trends, /api/v2/analytics/skill-saturation | | Market | /api/v2/market/skills, /api/v2/market/skills/:skill/trends | | Alerts | /api/v2/alerts/* | | Public matching | /api/public/v1/fit, /api/public/v1/companies/fit, /api/public/v1/mcp |

Frontend Pages

| Route | Page | |-------|------| | / | Jobs listing with search and filters | | /companies | Company profiles with drilldown | | /market | Market intelligence dashboard | | /market/sector/:slug | Sector deep-dive | | /market/salary/:category | Salary drilldown | | /frontier | Frontier AI hiring intelligence, official lab roles, evidence ledger, prep trails, and external review handoff | | /career | Career intelligence | | /talent | Talent pool analytics | | /app | Candidate job-search workspace | | /candidate | Candidate self-service portal | | /roles | Role taxonomy browser | | /saved | Saved items |

Project Structure

src/
  scraper/             # Playwright scrapers, pipeline, enrichment, publish (Python)
  feature_scoring/     # Feature/scoring CLI package
  enrichment_engine/   # Enrichment CLI package
  pages/, layouts/, content/   # Separate Astro marketing site sharing this src/ tree

reference/python-backend/
  server/              # Flask API (dev-time reference only — not deployed; see docs/ARCHITECTURE.md)

frontend/v2/
  src/
    pages/         # 20+ lazy-loaded route pages
    components/    # 400+ components across 30+ domains
    hooks/         # 90+ React Query hooks
    lib/           # API client, utilities, animation variants

docs/              # Architecture, product contract, guides
config/            # Scraper YAML configs
data/              # Local SQLite database (gitignored)

Documentation

| Document | Purpose | |----------|---------| | Product Contract | Mission, personas, canonical objects | | Master Plan | Execution roadmap | | Quick Start | 5-minute onboarding | | Workflows | Day-to-day commands | | Security | API security status |

License

Private. All rights reserved.