Installation Guide

How to Install and Run SkillBridge AI

Prerequisites

• Python 3.10 or newer. Grab it from python.org if you don't have it, and make sure you tick "Add Python to PATH" during install.
• Git (optional). You can just download and extract the project folder instead.
• Around 1 GB of free disk space for the ML models.
• An internet connection for the first run, since the embedding model and spaCy model download then.

Step-by-Step Setup

1. Open a terminal in the project folder. If you're cloning:

   git clone <repository-url> skillbridge
   cd skillbridge

2. Create a virtual environment. It keeps this project's packages away from everything else on your laptop.

   python -m venv venv

   Activate it. On Windows:

   venv\Scripts\activate

   On macOS or Linux:

   source venv/bin/activate

3. Install the dependencies. This takes a few minutes, so go refill your chai.

   pip install -r requirements.txt
   python -m spacy download en_core_web_sm

4. Create the database. By default the project uses config.settings.dev, which means SQLite and DEBUG on.

   python manage.py migrate
   python manage.py createsuperuser

   The second command is optional. You only need it if you want the admin site at /admin/.

5. Load the seed data. Order matters here, because internships depend on skills being loaded first.

   python manage.py seed_skills
   python manage.py seed_internships
   python manage.py seed_resources
   python manage.py build_embeddings

   These are safe to run again if something goes wrong halfway.

6. Start the server.

   python manage.py runserver

   Open http://127.0.0.1:8000/ in your browser.

7. Register an account at /accounts/register, finish onboarding (upload a resume from the samples/resumes folder if you want a quick test), then compute matches:

   python manage.py recompute_matches --all

Common Issues

• spaCy can't find en_core_web_sm. You probably ran the download outside your virtual environment. Activate the venv and run python -m spacy download en_core_web_sm again.
• The internship list says "Match not computed yet". Run python manage.py recompute_matches --all and refresh the page.
• First run is painfully slow. That's the all-MiniLM-L6-v2 model (about 90 MB) downloading. It's cached after this, so don't kill the process.
• Resume fields look wrong. Unusual headings or fancy layouts can trip the parser. Fix the values on the confirm page before saving.

How to Verify It's Working

Log in, open /dashboard, and you should see your profile summary, top internship matches and your top five skill gaps. Click any internship and check that it shows a percentage, five sub-scores and matched and missing skills. Then visit /gaps. If all three pages show data, you're done.

Want extra proof for your viva? Run python -m pytest tests/. The matching_engine and resume_parser tests are plain Python and need no database.