16 questions that I really like to ask during interviews for machine learning candidates.
These will help you practice, but more importantly, they will help you think and find ways to improve!
Let's do this! ☕️👇
These will help you practice, but more importantly, they will help you think and find ways to improve!
Let's do this! ☕️👇
1. How do you handle an imbalanced dataset and avoid the majority class completely overtaking the other?
2. How do you deal with out-of-distribution samples in a classification problem?
3. How would you design a system that minimizes Type II errors?
2. How do you deal with out-of-distribution samples in a classification problem?
3. How would you design a system that minimizes Type II errors?
4. Sometimes, the validation loss of your model is consistently lower than your training loss. Why could this be happening, and how can you fix it?
5. Explain what you would expect to see if we use a learning rate that's too large to train a neural network.
5. Explain what you would expect to see if we use a learning rate that's too large to train a neural network.
6. What are the advantages and disadvantages of using a single sample (batch = 1) on every iteration when training a neural network?
7. What's the process you follow to determine how many layers and nodes per layer you need for the neural network you are designing.
7. What's the process you follow to determine how many layers and nodes per layer you need for the neural network you are designing.
8. What do you think about directly using the model's accuracy as the loss function to train a neural network?
9. Batch normalization is a popular technique when training deep neural networks. Why do you think this is the case? What benefits does it provide?
9. Batch normalization is a popular technique when training deep neural networks. Why do you think this is the case? What benefits does it provide?
10. Why do you think combining a few weak learners into an ensemble would give us better results than any of the individual models alone?
11. Early stopping is a popular regularization technique. How does it work, and what are some of the triggers that you think could be useful?
11. Early stopping is a popular regularization technique. How does it work, and what are some of the triggers that you think could be useful?
12. Explain how one-hot-encoding the features of your dataset works. What are some disadvantages of using it?
13. Explain how Dropout works as a regularizer of your deep learning model.
14. How would you go about reducing the dimensionality of a dataset?
13. Explain how Dropout works as a regularizer of your deep learning model.
14. How would you go about reducing the dimensionality of a dataset?
15. Can you achieve translation invariance when processing images using a fully-connected neural network? Why is this the case?
16. Why are deep neural networks usually more powerful than shallower but wider networks?
16. Why are deep neural networks usually more powerful than shallower but wider networks?
I'll use these questions to guide some of the threads that I'll be posting in the coming days.
Stay tuned for that!
And don't forget to follow me so we can keep working on this machine learning stuff together!
🦕
Stay tuned for that!
And don't forget to follow me so we can keep working on this machine learning stuff together!
🦕
I usually try to work really hard in avoiding trivia questions. This is why I prefer to ask open-ended questions that allow candidates to show off their ability to reason.
I feel these are in that category. Do you feel differently about some of these?
I feel these are in that category. Do you feel differently about some of these?
https://twitter.com/GLegoy/status/1368286756664119297?s=20
Correct. An interview that's solely based on questions is a weak predictor of the future performance of that candidate.
The actual hands-on experience of the candidate plays an important role.
But this shouldn't preclude specific questions.
The actual hands-on experience of the candidate plays an important role.
But this shouldn't preclude specific questions.
https://twitter.com/zmnaka/status/1368328038472105984?s=20
Here is an explanation of what could be happening:
https://twitter.com/aureliengeron/status/1110839223878184960
https://twitter.com/bjornsing/status/1368236085487296512?s=20
Personally, I don't feel any of these questions are designed to trip candidates.
More importantly, they won't be asked looking to trip them, but instead, looking to start conversations that open a window to their experience.
More importantly, they won't be asked looking to trip them, but instead, looking to start conversations that open a window to their experience.
https://twitter.com/ThomasLWebb/status/1368505212827275265?s=20
• • •
Missing some Tweet in this thread? You can try to
force a refresh
