Since the beginning of this year, the author has been working in the fields of data science, machine learning and deep learning in India. During the thirty-four days of finding a job, he interviewed eight to ten companies, including start-ups, service-based companies, and product-based companies. The author hopes that his interview experience will provide useful information to job seekers and has written this article. I hope you can gain something after reading! First introduce yourself: I have more than 4 years of experience in machine learning (applications in speech analysis, text analysis, and image analysis). In general, I think most of the jobs in this field mainly include text analysis (natural language processing) and image analysis (computer vision). Few companies hire talent for voice or audio analysis. My current goal is to apply for a mid-to-high level position and lead a deep learning or machine learning team to do some interesting projects. Below are the questions I was asked during the application process and I hope I can help you. You mentioned in your resume that you have built a document mining system. What have you done? Can I use LDA technology to implement document clustering in topic modeling? Suppose you have hundreds of megabytes of data files, including PDF files, text files, images, scanned PDF files, etc. Please give a classification scheme. How do you read the contents of a scanned document in a scanned pdf file or image format? Why is Naive Bayes called "simple"? Please elaborate on the naive Bayes classifier. What is deep learning? What is the difference between deep learning and machine learning? Experience: In addition to this, the interviewer also asked some questions, but they all asked me, I don't know what answer he wants to hear. I have always wanted to talk in depth about technical issues, such as training a tesseract (an open source OCR engine developed by HP Labs maintained by Google) or a language model, but he doesn't seem to be interested. Maybe he just wants to hear some of the results that have been achieved or a good explanation, or some better solution. I feel that there is no difference between interviewing a newbie and interviewing an experienced professional. How to perform file clustering in unsupervised learning? How do I find files related to certain queries/searches? Explain the TF-IDF technology. According to my experience, TF-IDF technology does not work well in file classification or clustering. How will you improve? What is Long and Short Term Memory Neural Network (LSTM)? Explain how it works. What is the word2vec model? Explain the mutable and immutable objects in Python. What data structure have you used in Python? Experience: The entire interview process was all about text similarity, and I passed it. But this time there is still no deeper technical discussion. Perhaps the company has several small projects in the field of text analysis, and finally I got the company's offer. How to deal with multi-category classification problems with unbalanced datasets? How do you recognize a language from a text statement? How to represent pictographic characters in Chinese or Japanese? How to design a chat robot? (I don't have any idea, but I try to answer this question with intent and feedback based on TF-IDF similarity.) Can you use a circular neural network to design a chat robot to respond to the input questions in response and in response. Suppose you design a chat bot using a circular neural network or a long-short memory neural network on a Reddit dataset. It can provide 10 possible replies, how to choose the best response, or how to delete other replies? Explain how support vector machines (SVMs) learn nonlinear boundaries. Experience: There are still a few questions I can't remember. This is the first time I talked about the technical details in the interview. Then I got the offer from this company. What is precision and recall? Which one do you think is more important in medical diagnosis? Explain the accuracy and recall rate. How to plot the receiver operating characteristic curve (ROC curve)? What does the area under the ROC curve mean? How do I plot ROC curves for multi-category classification tasks? List other metrics for multi-category classification tasks. What is sensitivity (sensiTIvity) and specificity? What does “random†mean in a random forest? How to classify text? How do I know I have learned a text? Isn't it impossible to implement TF-IDF technology? (I replied to use the n-gram model (n=1, 2, 3, 4) and create a long count vector using TF-IDF technique) What else can you do with machine learning? (I recommend combining long- and short-term memory neural networks with word2vec, or a one-dimensional cyclic neural network combined with word2vec for classification. But the interviewer wants to improve the machine learning-based algorithm.) How does a neural network learn nonlinear shapes when a neural network consists of linear nodes? What is the reason for learning nonlinear boundaries? Experience: There are a few good questions I have not remembered. Although the entire interview process was very good, we did not agree on some issues. And during the interview, I found that as a startup, only 2-3 people are currently doing ML, DL and DS. Finally, I did not have an interview. What are the parameters when training a decision tree? What is the partitioning criterion for segmentation at a node in the decision tree? What is the formula for calculating the Gini coefficient? What is the formula for calculating entropy? How does the decision tree decide which feature must be split? How to use mathematical calculations to collect information? Briefly describe the advantages of random forests. Briefly describe the boosTIng algorithm. How does the gradient boost algorithm (gradient boosTIng) work? Briefly describe how the AdaBoost algorithm works. Which kernels are used in the SVM? What are the optimization techniques for SVM? How does SVM learn hyperplane? Discuss the details of its mathematical operations. Talk about unsupervised learning? What algorithms are there? How to define the value of K in the K-Means clustering algorithm? List the methods for defining K in the K-Means clustering algorithm in at least 3. What other clustering algorithms do you know other than this? Introduce the DB-SCAM algorithm. Briefly describe the working principle of Hierarchical Agglomera TIveclustering. Explain the principal component analysis algorithm (PCA) and briefly describe the mathematical steps of using the PCA algorithm. 20. What are the disadvantages of using the PCA algorithm? Talk about the working principle of convolutional neural networks? Details of its implementation details. Explain the back propagation in a convolutional neural network. How do you deploy a machine learning model? Most of the time we have to build a machine learning model from scratch in C++. Can you do that? Experience: I am interviewing a position at Amazon level 6. Their main focus is on algorithms and mathematics. But I didn't prepare for mathematics. I just talked about what I knew. I didn't discuss the details of mathematics in more detail, so the interviewer thought that I was not suitable for level 6. I believe that if you can remember the general representation of machine learning algorithms in mathematics, you can easily interview through Amazon technology. What is the scope of the Sigmoid function? Say the name of the package that scikit-learn can implement for logical regression. What is the mean and variance of the standard normal distribution? What data structure do you use in Python? What are the methods for text categorization? How would you do the classification? Explain the TF-IDF technology and its shortcomings, how to overcome the shortcomings of TF-IDF? What is Bigrams and Trigrams? Explain the TF-IDF technique of two-word collocation and three-word collocation with a textual statement. Give an example of what applications word2vec has. How to design a neural network? How to achieve "depth"? This is a basic neural network problem. Briefly describe how LSTM works. How does it remember the text? What is the Naive Bayes classifier? What is the probability of throwing 10 coins and 4 times positive? How do I get an index of an element in a Python list? What if you merge two pandas data sets? From the perspective of user behavior, you need to simulate a fraudulent activity. How would you solve this problem? This is probably an anomaly detection problem or a classification problem! Decision tree and random forest, which one do you prefer? What is the difference between logistic regression and random forest? Will you use a decision tree or a random forest to solve the classification problem? What are the advantages of random forests? Experience: I also got the offer of this company. In fact, I really like this technical exchange. Perhaps you will feel that these problems are the most fundamental issues in the field of machine learning and data science, but I feel that the interviewer may not be in this field, or that there is not much understanding of the development of this field. In the unbalanced data set, what model would you choose: random forest or Boosting? why? What do you know about Boosting technology? Using supervised learning to solve classification problems, which model do you choose? Suppose there are 40-50 categories! How do you use Ensemble technology? Briefly describe how the Support Vector Machine (SVM) works. What is Kernel? Briefly introduce. How to achieve nonlinear regression? What is Lasso regression and Ridge regression? Experience: To be honest, this interview was a bit so watery that I didn't take it seriously. But the question is very good. My interview position is to lead a team of fifteen or six people to do the project, after which is the manager interview and HR interview. In the end they gave me a job offer and a good salary. You mentioned the pronunciation recognition in the speech on your resume. Specifically, what is your implementation method? What is the Mel Frequency Cepstrum (MFCCs)? What is the Gaussian mixture model and how does it complete the clustering? How to maximize expectations? Talk about its implementation steps. How is the probability in the GMM model calculated? How do you perform MAP adjustments for GMM-UBM technology when performing pronunciation recognition? Talk about the I-vector technology you use. What are the main factors in analyzing context? What is the difference between JFA and I-vector? Why choose I-vector instead of JFA? Have you used PLDA I-vector technology? Have you read Baidu's Deep Speaker paper? If there are two models to choose from, what is the basis for your choice? (Exploring the technology of model selection) The mathematical working principle of Bayesian information metric (BIC) and Akaike information volume (AIC) is briefly described. What is the working principle of Bayesian information metrics and Akaike information? What should I do if the data in the MFCC feature vector matrix is ​​lost? How to perform speech recognition? What are the characteristics? Is your classifier a classifier for speech and music, or a classifier for speech and non-speech? How is deep neural network applied in speech analysis? Experience: Yes, you may be surprised at what these are. Coincidentally, the research areas of both of us are speech analysis (especially pronunciation recognition). So the entire interview process has been asking questions around speech analysis. Obviously, the interviewer was very professional and gave me a positive feedback. After that, the company gave me the job of an AI solution architect. During this entire job search process, I probably had conversations with 25-30 professionals. Here are my suggestions for readers and job seekers: The resume is very important. Be sure to include in your resume the projects you have participated in, the Kaggle competition, the MOOC course certificates or papers you have obtained. I received an interview with Amazon without any referral recommendation. Your resume is a weapon that impresses HR and the interviewer. Self-confidence and enthusiasm are half the battle. Be confident when attending an interview and show your enthusiasm to the interviewer (this is especially important when interviewing startups and service-based companies). Don't be too anxious to answer the questions asked by the interviewer. Take some time to organize the answers and answer them. If you don't understand the problem, be sure to ask the interviewer. There is also a calm time during the interview! Be sure to express yourself properly when explaining concepts. Give a few of the projects you have already implemented, and be familiar with the proficiency and projects you have mentioned in your resume. In most cases, the interviewer is looking for experienced technical talent in this field. If you are still a newcomer in this field, you can start with a project you have created when you create your resume. Your GitHub account is also very convincing. In addition, you can participate in the Kaggle competition and MOOC courses. In the face of the interviewer, be sure to be modest, pay attention to the interviewer's opinion, or you will be turned away. Sometimes, people who use R and Python will despise each other. You better not fall into this kind of argument, or you will be easily rejected. I personally think that both R and Python are tools for implementing logic and concepts. Finally, I wish you all a successful interview!
ZGAR bar 4000 Puffs
ZGAR electronic cigarette uses high-tech R&D, food grade disposable pod device and high-quality raw material. All package designs are Original IP. Our designer team is from Hong Kong. We have very high requirements for product quality, flavors taste and packaging design. The E-liquid is imported, materials are food grade, and assembly plant is medical-grade dust-free workshops.
Our products include disposable e-cigarettes, rechargeable e-cigarettes, rechargreable disposable vape pen, and various of flavors of cigarette cartridges. From 600puffs to 5000puffs, ZGAR bar Disposable offer high-tech R&D, E-cigarette improves battery capacity, We offer various of flavors and support customization. And printing designs can be customized. We have our own professional team and competitive quotations for any OEM or ODM works.
We supply OEM rechargeable disposable vape pen,OEM disposable electronic cigarette,ODM disposable vape pen,ODM disposable electronic cigarette,OEM/ODM vape pen e-cigarette,OEM/ODM atomizer device.
Disposable Vape, bar 4000puffs, ZGAR bar disposable, Disposable E-cigarette, OEM/ODM disposable vape pen atomizer Device E-cig ZGAR INTERNATIONAL(HK)CO., LIMITED , https://www.oemvape-pen.com
