AI Career Assistant

An AI-powered career guidance platform that helps students and professionals explore career paths, assess skills, find jobs, and receive personalized AI-driven recommendations.

Live demo: Deployed on Vercel — dark-themed, fully navigable, no account required.

---

What It Does

Feature	Status	Notes
Career Explorer (300+ O*NET careers)	Working	Static O*NET data, searchable
AI Career Assistant	Working	Powered by Claude via Vercel serverless API
Skills Analysis & Gap Identification	Working	Based on user profile
Job Search & Listings	Working	Curated job data
Career Pathways (structured roadmaps)	Working	Frontend → Full Stack → Data Science
Application Tracker	Working	Track applications through the hiring pipeline
Learning Hub	Working	Curated courses from top providers
Achievements System	Working	Gamified milestone tracking
Career Analytics	Working	Charts, metrics, skill progress
Profile Management	Working	Edit skills, title, location, education
Settings	Working	Theme, notifications, privacy, job preferences
Help Center	Working	FAQ + quick navigation
User Authentication	Not Implemented	Profile stored in localStorage
Real-time Job Scraping	Not Implemented	Backend not deployed to production
Persistent Cloud Storage	Not Implemented	localStorage only

---

Tech Stack

Frontend (deployed on Vercel)

• React 18.3 + TypeScript 5.8 + Vite 5.4
• Tailwind CSS 3.4 + shadcn/ui — component system
• Zustand 5 — state management
• React Query 5 — data fetching
• Framer Motion — animations
• Recharts — analytics charts
• React Router v6 — client-side routing

AI Integration

• Claude (claude-opus-4-6) via Vercel Edge Function at /api/chat
• Full user profile sent as system context on every request
• API key stored as Vercel environment variable (never exposed to browser)

Python Backend (local only)

• FastAPI — REST API (api_bridge.py)
• Gradio — parallel web UI (main_integrated.py)
• OpenAI GPT-4 + LangChain — AI modules
• SQLite — local data persistence
• Scikit-learn — TF-IDF and collaborative filtering
• BeautifulSoup4 + Selenium — O*NET scraping

The Python backend runs locally but is not deployed to production. The frontend uses static O*NET data and the Claude serverless function.

---

Quickstart

Frontend (the main product)

cd frontend
npm install
npm run dev
# Open http://localhost:8080

For the AI Assistant to work, create a .env.local file:

ANTHROPIC_API_KEY=your_key_here

Get your API key at console.anthropic.com.

Python Backend (optional, local only)

python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env      # Add your OPENAI_API_KEY
python api_bridge.py      # Starts FastAPI on port 8000

---

Deploying to Vercel

1. Push this repo to GitHub
2. Connect the frontend/ directory to Vercel (or configure rootDirectory: frontend in Vercel project settings)
3. Add ANTHROPIC_API_KEY as an Environment Variable in the Vercel dashboard
4. Deploy — the /api/chat serverless function is picked up automatically

The vercel.json in frontend/ handles SPA routing and API route exclusion.

---

Project Structure

AI-assistant-CSharpCorner/
├── frontend/                    # React/TypeScript frontend (main product)
│   ├── api/
│   │   └── chat.ts              # Vercel Edge Function — Claude AI proxy
│   ├── src/
│   │   ├── pages/               # All page components
│   │   │   ├── Dashboard.tsx
│   │   │   ├── AIAssistant.tsx  # Real Claude integration
│   │   │   ├── CareerExplorer.tsx
│   │   │   ├── JobSearch.tsx
│   │   │   ├── SkillsAnalysis.tsx
│   │   │   ├── Pathways.tsx
│   │   │   ├── Applications.tsx
│   │   │   ├── Learning.tsx
│   │   │   ├── Achievements.tsx
│   │   │   ├── Analytics.tsx
│   │   │   ├── Settings.tsx
│   │   │   ├── Profile.tsx
│   │   │   └── Help.tsx
│   │   ├── components/
│   │   │   ├── ErrorBoundary.tsx
│   │   │   ├── layout/          # Header, Sidebar, Layout
│   │   │   ├── careers/
│   │   │   ├── dashboard/
│   │   │   └── ui/              # shadcn/ui components
│   │   ├── services/
│   │   │   ├── api.ts           # API service (static data mode)
│   │   │   └── staticApi.ts     # O*NET static data service
│   │   ├── store/
│   │   │   └── useUserStore.ts  # Zustand user state
│   │   ├── data/
│   │   │   └── onetCareers.ts   # 300+ O*NET careers (static)
│   │   └── types/               # TypeScript type definitions
│   ├── vercel.json
│   └── package.json
│
├── modules/                     # Python backend modules
│   ├── career_roadmap_engine.py
│   ├── job_scraper.py
│   ├── skills_matcher.py
│   ├── recommendation_engine.py
│   ├── student_pathways.py
│   └── application_assistant.py
├── api_bridge.py                # FastAPI backend
├── main_integrated.py           # Gradio web UI
├── requirements.txt
└── config/system_config.yaml

---

AI Assistant Architecture

The AI Assistant uses a Vercel Edge Function (/api/chat) that:

1. Receives the conversation history and the user's full profile from the browser
2. Builds a personalized system prompt with the user's skills, title, experience, and interests
3. Calls the Anthropic Claude API (claude-opus-4-6)
4. Returns the response to the browser

The API key never touches the browser — it lives only in Vercel's environment variables.

Browser → POST /api/chat → Vercel Edge Function → Claude API → Response

---

Known Limitations

• No real authentication — user profile lives in localStorage. Data doesn't sync across devices.
• Job listings are curated mock data — not scraped live in production (backend not deployed).
• O*NET career data has quality gaps — skills/tasks fields are empty for most careers (scraper limitation).
• No code splitting for the 6,000-line career data file — impacts initial load time.
• SQLite backend not production-ready — would need PostgreSQL for concurrent users.

---

Roadmap

• Deploy FastAPI backend to Railway/Render
• Real-time job scraping via the Python backend
• User accounts and cross-device sync (Supabase or Firebase)
• Fix O*NET skills data quality (re-run scraper with corrected field extraction)
• Resume and cover letter generation (connect to Python ApplicationAssistant)
• Split onetCareers.ts into lazy-loaded chunks
• Migrate SQLite to PostgreSQL for backend persistence

---

Author

Rahul Mehta — GitHub

Originally developed for C# Corner. Powered by Claude (Anthropic) and O*NET data from the US Department of Labor.