The Library

The Library is the home of the app. Every interview Rubric knows about lives here, organized by category and ready to launch.

Anatomy of an interview

A single interview is a small package of three artifacts:

• Candidate board — an Excalidraw whiteboard the candidate sketches on during the session. It comes pre-loaded with a problem statement and (sometimes) a starter diagram.
• Interviewer guide — a private Markdown document with the full problem brief, evaluation criteria, follow-up questions, red flags to watch for, and hints you can offer if the candidate stalls.
• Rubric — a weighted YAML schema with three to five evaluation categories. Each category has six descriptors (one per score from 0 through 5).

Together those three files are everything you need to run, score, and report on a 30-minute round.

Categories

The library is organized into four top-level categories, each with its own letterpress accent color:

• AI / ML — model serving, content moderation, recommendation systems.
• Backend — APIs, queues, caches, classic distributed-system problems.
• Data — pipelines, warehouses, real-time analytics, schemas at scale.
• Infrastructure — deployment, observability, multi-region patterns.

Use the sidebar’s category filter to narrow the grid. The chip beside each category shows how many interviews are in it.

Browse, search, filter

The library grid supports several ways to find the right interview:

• Search — type any word from the title, description, or tags.
• Category filter — click a category chip to scope the grid.
• Difficulty filter — easy / medium / hard.
• Tag filter — domain tags (caching, streaming, multi-region, etc.) cluster related interviews.

The filters compose. “Backend + hard + caching” gives you the hardest backend caching problems Rubric knows.

The sidebar

The library sidebar shows three live counts:

• Active sessions — interviews you have in progress but haven’t ended. Click to resume.
• Past reports — finished interviews and their generated reports. Click to re-open.
• Help — opens this user guide.

The active-sessions list is the safety net for “the candidate’s video call dropped” or “I need to step away for ten minutes.” Sessions persist across app restarts.

The interview detail screen

Clicking any interview in the grid opens its detail screen. You see:

• Title, category, difficulty, duration.
• Target roles (e.g., “Staff Backend Engineer”, “Senior ML Engineer”).
• The description — what the problem is and what’s being evaluated.
• A preview of the rubric categories and their weights.
• A Start Session button.

Use the detail screen to brief yourself for two minutes before starting. The interviewer guide itself only appears once a session is open — the detail screen is a public-safe summary.

Where the library comes from

The interview catalog is bundled into the app — every install ships with the same starting set, ready to run offline.

The way to add your own interviews is the Generate screen, which uses your configured AI provider to author a complete interview package (board, guide, rubric) from a one-line role prompt. Generated interviews are added to your library right away and the source files are saved to your configured storage backend (see Storage Backends).

The bundled catalog itself is read-only — generated interviews live alongside it, not on top of it.