The job title written when submitting the resume is "Technical Research: Computer Vision". After the written test and personality test, I received an interview notice, which showed that it was developed in the background. I thought that the resume was not favored by the interviewer of the corresponding position, but was picked up by other interviewers. When I saw it, I fished it out. I was tangled on whether to go to the interview. In the end, I felt that I should go for it. Maybe the other party really needs to do visual.
4.16 On-site interview
Scan the QR code to sign in. The originally agreed time was 2:50 pm. I arrived earlier, at 1:50. After scanning the QR code to sign in, I received a prompt at 2:10 to go to the interviewer’s room.
The interviewer was in a standard room with a small table and two stools. I first introduced myself, and then the other party said that I was the SNG department of Tencent's headquarters in Shenzhen. I didn't know what department it was at the first time. I guess I was too nervous. In the Penguin eSports department, my first reaction was that it was Tencent's eSports platform. He said it was a live broadcast of the Penguin game, and then I realized it. I even watched the football match on it...
The interviewer asked me to introduce my papers, the first and the second, to tell him the difference between weakly supervised object detection and supervised object detection. It can be seen that he doesn't know much about vision, and then he said that they want to do something to filter the most concerned frames in the video. I thought it was key frame extraction, but it doesn't seem to be the case. I have to ask Huang Peiliang and Ye Senmao for more information. How do other directions in computer vision do.
I said that I use caffe and pytorch. He asked me to compare the two frameworks. I said that the caffe layer is more troublesome to write, and the pytorch is very simple to write the network structure, and it is more convenient to debug. The interviewer said that he thinks that there are many libraries in python that are more convenient.
Later, the interviewer asked about the network, programming framework, and database, but they didn't understand it... It's embarrassing to have an oral interview! Said that I am more worried about whether I can do anything else after I go in and finish the visual part. I feel cold when I hear this, and I will definitely not be able to do anything else.
At 7:00 p.m., I checked the results, "This position may not be suitable for you." As expected, I originally thought that I would go in first, but the senior brother in the laboratory said that if I changed my career after graduation, I felt a little bit of a loss, and these things would be in vain. learned. After thinking about it carefully, it is indeed the case. The goal is to choose the corresponding visual direction to work better, otherwise, not only the previous white learning, but also learning other things from scratch.
In short, Tencent's intern recruitment has ended, and Ali's has also ended. Baidu's written test will be held on the 18th, and we will wait for the results.