Knowledge base

In short: the knowledge base is the context the LLM relies on when building answers. In Mockingbird it's usually a PDF résumé and technology notes. The file is processed locally and never sent anywhere.

Loading a résumé

1. Open the Résumé tab.
2. Click Upload PDF and pick a file.
3. Wait for indexing — the app shows the number of chunks.

The résumé is processed locally via pypdf — the file is never uploaded anywhere.

How it affects answers

When the interviewer asks a question, relevant knowledge base chunks are mixed into the LLM context. The more precise and structured your résumé, the more specific the answers.

• List technologies and projects with details (what you did, with which tools).
• Split it into sections: Experience, Stack, Projects.
• Avoid scanned images without a text layer.

Refreshing the knowledge

To replace the context — upload a new PDF again: the old chunks are re-indexed.