Interview Steps:

• Interview with HR
• Take-home test
• Debrief of the take-home test
• Discussion with the CTO
• Offer

The process is well organized and simple, but very deceptive. I made it clear what the minimum salary I was expecting was. HR told me it was possible. Even with a "very good" technical exercise, the maximum salary offered was 20% lower than my current one, which I shared with them at the beginning of the process.

Stop wasting people's time! If you can't match their salary, just say it at the beginning of the process. During the technical debrief, even when their technical team said they "learned a lot" from my solution, what else do you want people to do?!

For this exercise, you have to:

1. Pick one of the datasets from https://huggingface.co/datasets.
2. Implement a web API to launch queries on the dataset's contents (e.g., full-text search among the contents, or aggregation over the values of a column).
3. Pick any tool, language, or framework you want, but explain why.
4. Choose a query language, explain why it's the most appropriate, and generate minimal "Getting Started" documentation.
5. Deploy your API in the cloud. Pick any cloud provider and hardware. We can reimburse any hosting costs you may incur.

Your deployment should be reproducible and somewhat scalable.