Uber Data Scientist Mock Interview: Ride Requests Model

  Рет қаралды 29,945

Jay Feng

Jay Feng

Күн бұрын

Пікірлер: 16
@clapdrix72
@clapdrix72 2 жыл бұрын
Sorry but this is horrendous. I was wondering why he seemed like he was stalling by over-focusing on the features until he said they could use a multi class model to bin the likelihood to accept at different levels, at which point it became clear... If I were interviewing him for a DS position at my company, I'd thank him for his time right after that he said that.
@bibiworm
@bibiworm 3 жыл бұрын
The interviewee obviously was not familiar with the product or multi sided market. As far as I know, drivers are not shown the address or the neighborhood of the rider before they accept the trip just so they won’t cherry pick. So the interviewer finished feature selection. But I like how comfortable both parties were here.
@pro100olga
@pro100olga 3 жыл бұрын
As I understand, they know the pick-up address, but not the destination - due to the reason you have mentioned
@datahat642
@datahat642 3 жыл бұрын
Very interesting case study with lots of details. I would like to add up two key features: Average number of trips taken by the drive which could be further subdivided as average number during weekdays and average number during weekends and how close the drive is to that value. 2. Is there an additional incentive which will be closer to the driver if they take the ride, say for example bonus after completing certain number of trips.
@kumarcal
@kumarcal 2 жыл бұрын
Thanks for the video!!! Some feedback: - Didn't go into why the model is needed or what it's used for - Did not exactly define how this model would be evaluated? Aka how we define success - Missed subject of how we would run experiments for this model and also retrain if needed (related to how the model would be deployed). ^^I'm a senior ML/NLP engineer trying to figure how to do ML architecture interviews :D
@shairuno
@shairuno 2 жыл бұрын
THe interviewee is PM for a reason. He could come up with some novel use cases and potential product features.
@saradhisomarouthu376
@saradhisomarouthu376 4 ай бұрын
The Driver's incentive i.e. How much does the driver earn from the ride would be the most important factor, how was that not given importance.
@bibiworm
@bibiworm 3 жыл бұрын
Did he sync iPad with Mac and then write on Ipad? Thanks.
@varunmanjunath6204
@varunmanjunath6204 3 жыл бұрын
Way easier than ds and algo
@spire52
@spire52 4 жыл бұрын
Thank you ! really informative
@kaiyanqiu6366
@kaiyanqiu6366 4 жыл бұрын
Thanks so much for sharing!!!
@spiritunchecked2281
@spiritunchecked2281 4 жыл бұрын
Awesome effort Jay!
Остановили аттракцион из-за дочки!
00:42
Victoria Portfolio
Рет қаралды 3,8 МЛН
when you have plan B 😂
00:11
Andrey Grechka
Рет қаралды 67 МЛН
The joker favorite#joker  #shorts
00:15
Untitled Joker
Рет қаралды 30 МЛН
Please Help This Poor Boy 🙏
00:40
Alan Chikin Chow
Рет қаралды 22 МЛН
A Day in Life as a Google Data Scientist (Analytics)
6:43
Sundas Khalid
Рет қаралды 403 М.
Spotify Data Scientist Business Case Interview
28:40
Jay Feng
Рет қаралды 16 М.
PM Mock Interview: How to improve Uber ETA?
18:02
JJ never sleeps
Рет қаралды 1 М.
Amazon Data Scientist Mock Interview - AB Testing
16:15
DataInterview
Рет қаралды 27 М.
Uber Product Manager Mock Interview: Estimate Drivers in SF
13:38
Остановили аттракцион из-за дочки!
00:42
Victoria Portfolio
Рет қаралды 3,8 МЛН