The questions asked by HR are as follows:
1. Tell me about the algorithm used in the project and how effective it is. I will talk about time-domain behavior detection, time-series convolution, graph convolution, target detection and so on.
2. Asked about the details of the behavior detection algorithm. Such as data preprocessing method, time domain positioning method, time domain positioning error evaluation method, recall rate definition and so on.
3. What framework is used by the target detection algorithm. yolov3, talk about its advantages, how to locate small goals, and the definition of loss.
4. What kind of code do I often write? I just said to build a special target detection framework from scratch to adapt to the needs of the project, and then asked some algorithm improvements.
5. Data structure, traversal and reversal of binary tree.
6. Comparison of commonly used machine learning algorithms, such as svm, xgboost, gbdt, and clustering.
7. Handwriting sorting algorithm (quick sorting).
8. Hand tear code:
(1) Given an array, output all combinations of any three numbers and 0.
(2) Return the longest common subsequence of two strings. The subsequence is the sequence obtained by deleting some characters from the original string without changing the relative order of the characters.
Interview experience:
After a few days, he was rejected. . .
1. My basic situation: a
general undergraduate degree, a master's degree in postgraduate research, and a major in mechanical engineering. For the paper, there are two EI papers (one has been published, one is under review), one is the core science and technology, one invention patent, and two soft works. The instructor's key project leader, the leader of the research group.
2. Interview situation: The interview
questions are basically answered, and the interviewer mentioned a point, saying that I am engaged in deep learning, and is it not a good match for the machine learning position (ant mainly doing big data prediction)?
3. Summary of the reasons for personal failure:
a. Low academic qualifications and non-subjects (Generally Ali requires 985 or 211 computer majors);
b. The delivery position is not suitable;
c. My skills are too good, and there is no machine learning related top. Meeting thesis;
d. The principle knowledge is not solid, and the previous period has not been memorized seriously.