J'ai postulé en ligne. Le processus a pris 2 semaines. J'ai passé un entretien chez PathAI (Boston, MA) en nov. 2019
Entretien
There are 3 interviews: one online coding challenge, one machine learning basic phone interview, and one onsite interview about both software engineer and machine learning stuffs.
I first started with the online coding challenge. There were 2 parts: multiple choice questions about basic classical machine learning and one coding problem. The multiple questions weren't very hard. The coding problem needs a bit of effort, but it was a standard challenge anyone can find on leetcode/hackkerrank.
The next interview was a 30-40min phone interview about deep learning. Basically, they will test your knowledge about neural network from activation layers. batchnorm, initialization, etc..
Lastly, I was invited to have a 4 hrs onsite interview with the ML members of the company and CTO. There were 5 small interviews sections: classical ML, math & probability, algorithm, deep learning, and behaviors question. It was definitely a long and exhausted interview, but overall, I had positive experience.
Two coding problem. Easy, array focused and debugging. Two ML screens - one on research and the other on systems. New grad can crack after 1 month of prep. ML was around regularization and second order optimization.
J'ai postulé en ligne. Le processus a pris 1 semaine. J'ai passé un entretien chez PathAI en janv. 2022
Entretien
Applied through LinkedIn. The recruiter reached out over email to do a online HackerRank assessment of a custom machine learning problem for AI and it was timed at 60 minutes.
Questions d'entretien [1]
Question 1
The problem was to calculate K nearest neighbors for each of their 2D arrays given another array. Using python libraries was allowed.
J'ai postulé en ligne. J'ai passé un entretien chez PathAI
Entretien
Applied online, directly got hacker rank link. online assessment consists of Multiple Choice Questions + Programming question.
Multiple Choice Questions are basic ML +CNN + Math questions. about 90 mins total
Questions d'entretien [1]
Question 1
Calculate Parameters for given CNN network given the network architecture.
Other ML+CNN questions.
Programming question: Jaccard Coefficient. The examples mentioned were misleading