Cityblock Tech Interview — After Hours Support

A monorepo for the System Design Interview: a GraphQL web app for managing nurse availability, plus a Python ETL pipeline that ingests partner clinic feeds into the same users table.

The candidate-facing PRD lives in PRD.md (also shared as a Google Doc at the start of the interview).

The scaffold ships the legacy User model ({ availableDays: string[] } jsonb) end-to-end. Extending it is the exercise.

What's here

Component	Path	What it does
Web app	packages/	React UI + GraphQL server for viewing and editing nurse availability
ETL pipeline	etl/	Ingests partner clinic CSV/JSON feeds into the shared users table

Both components read and write the same file-backed SQLite database (data/app.sqlite by default), so rows ingested by the ETL show up immediately in the running app.

Quick start

Web app — Node 22, pnpm 10:

pnpm install
pnpm dev

Web: http://localhost:5173 · GraphQL: http://localhost:4000/graphql

ETL — Python 3.12+, uv:

cd etl
uv sync
uv run etl

See the component READMEs for full setup, configuration, and layout details:

• packages/README.md — web app and GraphQL server
• etl/README.md — partner clinic feed ingestion pipeline

Layout

packages/
  server/   GraphQL Yoga + Knex (better-sqlite3, file-backed)
  web/      Vite + React 18 + Apollo Client
etl/        Standalone Python pipeline that ingests partner clinic feeds into `users`

Shared configuration

Env var	Default	What it does
PORT	4000	GraphQL server port
DB_FILE	data/app.sqlite	SQLite file path shared by the server and ETL. Set DB_FILE=:memory: for an ephemeral database.