Objective

Build a backend agent that can understand and respond to natural language weather-related questions using live data from weather.com.

You’re free to decide the interface (CLI, API, bot, web app, etc.) and architecture. The goal is to evaluate how you think about product and technical decisions when given a broad problem statement.

Requirements

• Pull live weather data
• Support natural language queries
• Build a utils for a google calendar integration that we can use to extend this product (to be used during the interview)
• The agent should work end-to-end and be testable.
• You can use your own tech stack; however, note that we work with Node JS, Python and Typescript

Deliverables

• A live, working version of the agent that we can test.
• A Loom video ( < 2 min) showing:
  • How the product works
  • How to interact with it
  • A quick tour of the codebase and key architectural decisions
• A README.md including:
  • Setup instructions
  • Description of your approach
  • Key product and engineering decisions

Next steps

• When ready, send your Loom link, test link, and availability for a 1-hour follow-up to [email protected].
• In the technical interview, we’ll:
  • Walk through your solution together
  • Collaboratively extend or modify the product (please have your dev environment ready)