A semantic search engine and AI assistant for CV skills built on Cloudflare's edge platform. The assistant provides outcome-driven, recruiter-ready answers that focus on measurable business impact rather than generic claims.

• Outcome-Driven Responses: Every answer follows the template: Skill → Context → Action → Effect → Outcome → Project
• Semantic Search: Vector-based similarity search using D1 + Workers AI embeddings
• Edge-Native: Runs on Cloudflare Workers with D1 database and Vectorize
• AI-Powered Replies: LLM-generated responses via Workers AI (@cf/meta/llama-3.2-3b-instruct)
• No Fluff: Avoids recruiter buzzwords, focuses on measurable results
• Fully Automated: One-command deployment with PowerShell automation
• Turnstile Protected: Cloudflare Turnstile prevents bot abuse

# Full automated deployment (recommended)
npm run deploy:full

This single command:

1. ✅ Builds TypeScript
2. ✅ Deploys to Cloudflare Workers
3. ✅ Checks/applies database migrations
4. ✅ Seeds data if needed
5. ✅ Generates embeddings (64 records)
6. ✅ Runs health checks
7. ✅ Takes ~2-3 minutes total

# Install dependencies
npm install

# Build TypeScript
npm run build

# Deploy worker
npm run deploy

# Apply schema migration
wrangler d1 execute cv_assistant_db --remote --file=migrations/003_add_outcome_fields.sql

# Seed database
wrangler d1 execute cv_assistant_db --remote --file=migrations/002_seed_data_tech_only.sql

# Index vectors
npm run index:remote
wrangler d1 execute cv-assistant-db --file=migrations/001_initial_schema.sql

# Seed data
wrangler d1 execute cv-assistant-db --file=migrations/002_seed_data.sql

# Deploy
npm run deploy

curl "https://your-worker.workers.dev/query?q=Tell me about your microservices experience"

Response:

{
  "query": "Tell me about your microservices experience",
  "results": [...],
  "assistantReply": "With 5+ years of advanced experience in Full‑Stack Service Decomposition at CCHQ, I broke down monolithic applications into modular services. This enabled teams to deploy independently, cutting release cycles from weeks to days.",
  "source": "d1-vectors",
  "timestamp": "2025-01-15T10:30:00Z"
}

"I'm an exceptional engineer with a proven track record in microservices."

"With 5+ years of advanced experience in Full-Stack Service Decomposition at CCHQ, I broke down monolithic applications into modular services, cutting release cycles from weeks to days."

The technology table includes outcome-driven fields:

CREATE TABLE technology (
  id INTEGER PRIMARY KEY,
  name TEXT NOT NULL,
  experience_years INTEGER,
  level TEXT,
  summary TEXT,
  action TEXT,              -- What was done
  effect TEXT,              -- Technical effect
  outcome TEXT,             -- Business result
  related_project TEXT,     -- Project context
  -- ... other fields
);

Set environment variables in wrangler.toml:

[vars]
AI_REPLY_ENABLED = "true"

• GET / or GET /health - Health check
• POST /query or GET /query?q=<query> - Semantic search with AI reply
• POST /index - Index all skills into vectors

• Cloudflare Workers - Edge compute
• D1 - SQLite-based database
• Workers AI - Embedding generation and LLM responses
• Vectorize - Vector similarity search (optional)
• TypeScript - Type-safe development
• Vitest - Unit testing framework

Run the test suite:

# Run all tests
npm test

# Run tests in watch mode
npm run test:watch

# Run with coverage
npm run test:coverage

Test Coverage:

• Service layer tests (QuestionValidator, ResponseValidator, ProjectDetector)
• Error handling and type utilities
• 156 test cases with comprehensive edge case coverage

• Architecture
• Database Deployment
• Database Seeding

• Re-seeding Guide - How to update AI data
• Schema Migration Guide - ⚠️ Read this before adding new fields
• Re-seed Solution Summary - Quick reference

Built with ❤️ on Cloudflare's edge platform