Fiz uma entrevista na empresa Unit 410 (Austin, TX).
Entrevista
three rounds, code pairing, hr, and work trial. work trial is paid. pretty standard, nothing crazy. code problem is a pairing exercise presumably designed to mimic realworld a problem and work scenario.
Candidatei-me por meio de recrutador(a). O processo levou 5 semanas. Fui entrevistado pela Unit 410 (New York, NY) em jun. de 2024
Entrevista
The interview process began with speaking with the HR head, who walked me through the process in a very detailed manner. I've never interviewed at a company that helped candidates with this level of detail and context. There were materials for candidates to review about the company to help answer FAQs.
The next steps in the process were to meet various members of the team, who were across-the-board positive on the company. I received a very good sense of what the job would entail, the current challenges, and the incentive structure.
While the process didn't land in an offer, I only had good things to say about my experience and would consider the company again in the future if the job opportunity better aligned.
Perguntas de entrevista [1]
Pergunta 1
They asked me to walk through a blockchain network that I was bullish on to get a sense of my familiarity with general network concepts and specific network differentiators.
Candidatei-me online. O processo levou 2 semanas. Fui entrevistado pela Unit 410 em dez. de 2023
Entrevista
The interview was relatively straightforward.
The first part was a standard discussion with a recruiter.
The second part was a coding challenge. The challenge wasn't terribly difficult. However, the interviewer I had showed up late, kept his camera off the entire time, and clearly had young kids running around in the background and was distracted. It was one of the most uncomfortable interviews I've ever had and honestly would have been better had I just been solving the problem on a site that records the screen.
Perguntas de entrevista [1]
Pergunta 1
The question was to take a stream of buy/sell requests for an asset and put together an order book.
The structure of the data was this:
{ "type": "buy", "price": 10.0, "quantity": 0.3 }
{ "type": "buy", "price": 9.5, "quantity": 2.9 }
{ "type": "buy", "price": 9.0, "quantity": 0.1 }
{ "type": "buy", "price": 8.5, "quantity": 1.2 }
{ "type": "sell", "price": 10.5, "quantity": 2.4 }
{ "type": "sell", "price": 11.0, "quantity": 1.4 }
{ "type": "sell", "price": 11.5, "quantity": 1.3 }
{ "type": "sell", "price": 12.0, "quantity": 0.9 }
The question was:
Write some code to operate a fake order book that accepts buy and sell orders. Populate the orderbook with some buy and sell orders using price increments of 0.5 and pseudo-random values for the quantity. The end result should look similar to the example(s) above.
It should:
1. Update the state of the orderbook as appropriate.
2. Produce trades as appropriate.
