Candidatei-me online. O processo levou 2 semanas. Fui entrevistado pela Hex PixelPhant (Udaipur, Rajasthan) em dez. de 2023
Entrevista
Two rounds of intro call. One with the HR and one with the CTO. It was a moderately difficulty assessment. I had to train a segmentation model for an image segmentation task. They had provided a dataset. I chose the pre-trained DeepLab V3. Following this, I had one more round and had to demonstrate an inference on an image of their choice. Finally had an HR round.
Perguntas de entrevista [1]
Pergunta 1
How did you handle multiple masks representing multiple individuals in an image. Which loss function did you use and why? Run an inference on an image provided by them.