senior solutions engineer interview questions shared by candidates
If you're only given the data without knowing anything about the dataset, how would build a machine learning model from it?
The first step is to review the distributions, understand the data types, visualize, etc. This is a pretty straightforward Data Science question
If all the variables are numerical: 4 steps: 1. Matrix Scatterplot, 2. Center and Scale the features , 3. Either dimension reduction (PCA or T-SNET depending on the data) and 4. cluster . One can also use an encoder to reduce the dimension space. If all the variables are binary (one-hot encoder) you can just use Boltzmann Machine to reduce your space and find clusters. However, it all depends on the characteristics of the data, I found out while this technic may work in most cases, I still stumbled into a case where such technic failed.
See Interview Questions for Similar Jobs
- Mechanical Engineer
- Registered Nurse
- Sales Associate
- Pharmacy Technician
- Graphic Designer
- Social Worker