AI-powered job compatibility analyzer built at the NemoClaw NVIDIA x ASUS Hackathon @ UC Santa Cruz (May 2026).

Fill out your profile once. A background agent continuously fetches live job listings from Adzuna, scores each one against your experience using NVIDIA Nemotron, and surfaces your top matches ranked by fit — no manual searching required.

agent-app/
├── backend/
│   ├── agent/
│   │   ├── analyzer.py   ← Nemotron batch compatibility scoring
│   │   ├── loop.py       ← background polling loop
│   │   └── tools.py
│   ├── main.py           ← FastAPI app, NDJSON streaming
│   ├── jobs.py           ← Adzuna fetching + Supabase storage
│   ├── profile.py        ← user profile API
│   └── auth.py
└── frontend/
    └── src/
        ├── pages/        ← Profile, Home (top matches), Jobs
        ├── components/
        └── App.jsx

1. Profile — enter your skills, experience, projects, and job preferences.
2. Fetch — every 30 minutes (or on demand) the backend pulls fresh listings from Adzuna based on your preferred roles and locations.
3. Analyze — new listings are scored in batches of 8 using a single Nemotron LLM call per batch, producing a compatibility score (0–100), reasoning, and a ranked breakdown of your experiences by relevance to each job.
4. Match — the Home feed shows your top matches sorted by score.

You'll need two terminals and accounts for Supabase, Adzuna, and NVIDIA NIM.

cd backend
python -m venv .venv
source .venv/bin/activate    # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env         # fill in keys (see below)
uvicorn main:app --reload --port 8000

Required env vars in .env:

SUPABASE_URL=
SUPABASE_KEY=
NVIDIA_API_KEY=
NVIDIA_BASE_URL=https://integrate.api.nvidia.com/v1
NEMOTRON_MODEL=nvidia/nvidia-nemotron-nano-9b-v2
ADZUNA_APP_ID=
ADZUNA_APP_KEY=

cd frontend
npm install
npm run dev

Open http://localhost:5173. Vite proxies /jobs, /profile, and /run to the FastAPI server on :8000.

• FastAPI — async-native, painless NDJSON streaming
• NVIDIA Nemotron via NIM — batch LLM calls for job compatibility scoring
• Adzuna API — live job listings
• Supabase — auth + database
• React + Vite — fast dev loop

Built in 24 hours at the NemoClaw NVIDIA x ASUS Hackathon @ UC Santa Cruz — the premier physical AI hackathon on the West Coast, hosted at Kresge College, May 15–16 2026. Sponsored by NVIDIA, ASUS, and the Baskin School of Engineering.