What have I learned from interviewing AI companies over the past eight months?

This is the most important step. Getting a recruiter to read their profile through a mountain of resumes is the hardest part for candidates. In general, this can be broken down into three key steps:

First of all, your resume should be concise. This guide (https://career-resource-center.udacity.com/resume/resume-revamp) are available upon Udacity to sort out the resume. This guide contains a lot of things I want to say, and I’ve been using it as a resume guide for myself. When it comes to resume templates, Overleaf (http://overleaf.com) offers some pretty good ones, such as:

As you can see, the resume is placed on a single page. However, if your resume has too much content, the template above may not fit. A modified multi-page template can be found here (https://latexresu.me).

The next most important thing is how to present your Github profile. Many people underestimate Github simply because it doesn’t have a “who’s seen your profile” option.

Recruiters will look at what you put on Github because that’s the only way they can verify what you mention on your resume. People use a lot of buzzwords in their profiles, causing a lot of distraction for recruiters. For data science in particular, open source plays an important role in most tools, implementations of various algorithms, learning resources, etc., all of which are open source. Candidates should at least:

• If you don’t already have a Github account, create one.

• Create a repository for each completed project.

• Add documentation and provide explicit instructions on how to run the code.

• Add documentation for each file that explains what each function does and what each parameter means.

The next step, which most people don’t have, is to create a personal website to showcase your experiences and projects. This step can show that a person really wants to get into the field. In addition, resumes often don’t show everything and often miss a lot of details.

If you wish, you can show more details on the site, such as a visual presentation of a project or idea. It’s really easy to create a site like this, because there are many free platforms that offer drag-and-drop features that make this task very easy. I personally use Weebly (http://weebly.com), an already widely used tool. Here is my personal website:

Finally, a lot of recruiters and startups are using LinkedIn as their recruiting platform, and a lot of good jobs are being posted on it.

In addition to recruiters, headhunters looking for candidates for high-impact jobs are also active. So if you can get their attention, you may have a good chance.

In addition, it’s important to keep your profile simple so that people will want to connect with you. Search tools are an important part of LinkedIn, and in order to get the most out of search tools, you must use relevant keywords in your profile. Also, be sure to ask people who have worked with you to add comments about their experiences working with you. All of this increases your chances of getting noticed.

These things may seem like a lot, but they don’t have to be done in a day or even a week or a month. It’s a process, and there’s no end. It definitely takes a while to get everything ready in the beginning, but once it’s ready, it’s just a matter of updating it. You’ll find it easy, and you can sell yourself anywhere, anytime, because you already know yourself so well.

I’ve read a lot of resumes, and I’ve seen a lot of people make this mistake. In my opinion, it’s better to figure out what you’re really interested in and what you want to do first and then look for related opportunities, rather than the other way around. The fact that AI talent is in short supply offers a lot of opportunities. The process of preparing your resume will help you get a complete picture of yourself and help you make better decisions. In addition, there is no need to prepare answers to various questions that may be asked during the interview. Because when you’re talking about something you really care about, these answers come naturally.

If you don’t talk to people, you miss out on a lot of opportunities. Stay in touch with people on a daily basis, if not in person, then at least on LinkedIn, and over time you’ll have a huge network. Networking is not about asking people to make recommendations for you.

I made this mistake a lot in the beginning, too, until I came across an article by Mark Meloon where he talked about the importance of making real connections with people, starting with providing them with help. Another important aspect is showing what you know. For example, if you’re good at something, share it with people on Facebook and LinkedIn. This will not only help others, but also help yourself.

Once you build a good enough network, your visibility costs more. A comment from someone in your network about what you share could help you reach a wider audience, including professionals who may be looking for something like you.

To avoid any confusion, I’m going to make this list alphabetically. However, I put an asterisk in front of some companies’ names to indicate that I personally recommend them. Some companies have more than an asterisk because they are better at being human.

• Adobe Research

• ⭐ AllinCall (founded by IIT Bombay alumni)

• ⭐ amazon

• Arya.ai

• ⭐ Element. Ai

• ⭐ Facebook AI Research Institute

• ⭐ Fracal Analytics (acquired Cuddle.ai and ⭐⭐ Qure.ai)

• ⭐⭐ Google (Brain/DeepMind/X)

• Goldman sachs

• Haptik.ai

• ⭐⭐HyperVerge, founded by IIT Madras alumni, is dedicated to developing AI solutions to real world problems for clients around the world. These founders started the Computer Vision Group at IIT Madras.

• IBM research institute

• ⭐ Intel Artificial Intelligence Lab (Reinforcement Learning)

• ⭐⭐Jasmine. Ai, founded by IIT Madras alumni who also earned their PHDS at the University of Michigan, is researching intelligent conversation. In addition, they are flush with cash and want someone to join their Bangalore office soon.

• Jp Morgan chase

• ⭐ Microsoft Research

• MuSigma

• Next Education

• niki.ai

• ⭐Niramai, a former Xerox employee, uses thermal imaging to detect breast cancer at an early stage.

• Ola

• ⭐ OpenAI

• ⭐ PathAI

• Predible Health

• qualcomm

• ⭐ SalesForce

• Samsung research

• ⭐ SigTuple

• ⭐Suki, AI-powered voice assistant for doctors. It has also recently raised substantial funds and may soon open an office in India.

• ⭐Swayatt Robotics, dedicated to developing self-driving cars for India.

• ⭐⭐Wadhwani AI, funded by billionaires Romesh Wadhwani and Sunil Wadhwani, aims to be the first AI organization working for social welfare.

• ⭐ Uber AI Lab and Advanced Technology Group: THE AI Residency Program

• ⭐Umbo CV, Computer vision based security technology

• Uncanny Vision

• Zendrive

The interview will start from the moment you enter the room, and a lot can happen before you start introducing yourself: your body language and smile on your face are very important, especially when you’re interviewing with a startup, because culture fit is something they’re very focused on. Understand that the interviewer is a stranger to you, but you are also a stranger to him or her. So, they’re probably just as nervous as you are.

The interview isn’t just a conversation between you and the interviewer. Both sides are looking for a mutual fit: you’re looking for a great place to work, and the interviewer is looking for a great person to work with. So make sure you feel good about yourself and make the other person feel comfortable at the beginning of the conversation. The easiest way is to smile.

There are two main types of interview, one is when the interviewer comes to you with prepared questions, regardless of what your resume says, and the other is when the interview is based on your resume. I’ll start with the second.

This interview usually starts with “Can you tell me something about yourself?” To begin with. There are two big no-no’s in answering this question: talking about your college gpa or going into too much detail about projects you’ve worked on. The ideal way to respond should be to explain what you’ve done so far in about a minute or two. You can talk about your hobbies, such as reading, sports, meditation, etc. The interviewer will use what you have said as a clue to the next question and then move on to the technical part of the interview. This part of the interview is to check that what you have written on your resume is true.

There are a lot of “what if” questions, such as what would happen if “X” was used instead of “Y”. When answering these questions, it’s important to know what trade-offs are commonly made during implementation. For example, if the interviewer says that using a more complex model will give better results, you can say that because there is less data available, it will lead to overfitting. In an interview, I was asked to design an algorithm for a real case. I’ve found that interviewers are satisfied when I follow this process:

Question > Previous one or two methods > My method > Results > Intuition

The other is just to test your knowledge of the basics. These questions won’t be too difficult, but they will certainly cover the basics you should know, such as linear algebra, probability theory, statistics, optimization, machine learning, and deep learning. The amount of time you spend answering these questions is critical. Since these cover the basics, they expect you to be able to say the answer right away, so be prepared.

Throughout the interview process, it’s important to be confident about what you know and honest about what you don’t know. If you don’t know the answer to a question, explain the situation first, instead of being “MMMM”. If some concept is really important but you’re having trouble answering it, the interviewer will usually be happy to give you a hint or guide you to the right solution. If you can follow their cues and come up with the right solution, you’ll score points. Try not to be nervous, and the best way to avoid this is still to smile.

At the end of the interview, the interviewer will ask you if you have any questions. You think the interview is over and there’s nothing left to ask. I know many people who have been rejected for being wrong on this issue. As I mentioned earlier, interviews are not one-sided, as you are being interviewed, you are also looking for ways to fit in with the company itself. So, if you’re serious about joining a company, you’re sure to have a lot of questions about the company’s work culture, or about the position you’re applying for. You want to make sure you give the interviewer the impression that you are genuinely interested in being part of their team. At the end of the interview, I ask the interviewer what they would like me to improve on. It helped me tremendously, and I incorporated every feedback they gave me into my daily life.

That’s it. In my experience, if you are honest with yourself, competent, genuinely care about the company you are interviewing with, and have the right mindset, you are well on your way to success and will receive an offer soon.

We live in an era of opportunity and you just have to be the best and you’ll find a way to cash in. As Gary Vaynerchuk puts it:

It’s a great time to work in AI, and if you’re really passionate about it, there’s a lot you can do with AI. We complain all the time about the problems around us, and it’s unprecedented that ordinary people like us can actually do something about them instead of complaining. Jeffrey Hammerbacher (Cloudera founder) once said:

It’s too bad my generation is trying to figure out how to get people to click on ads.

We can do a lot of things with AI that we never imagined. There are a lot of very challenging problems that require smart people like you to solve. You can change people’s lives for the better.

An interview for any data science-related job will consist of questions in four categories: computer science, mathematics, statistics, and machine learning.

Algorithms and data structures:

• InterviewBit (https://www.interviewbit.com)

• NPTEL IIT Delhi lecture on YouTube (https://www.youtube.com/playlist?list=PLBF3763AF2E1C572F)

Operating system:

• What Software Developers need to know 10 Operating system concepts (https://medium.com/the-aspiring-programmer-journal/the-10-operating-system-concepts-software-developers-need-to- Remember – 480 d0734d710)

• “Operating system concepts.” chapters 3,4,5, and 7.

• GeeksForGeeks (https://www.geeksforgeeks.org/operating-systems) of the operating system

Object-oriented programming: You may be asked how to design a system, such as a railway ticketing system. So, you need to figure out what the interviewer’s requirements are, what classes you need to create, what variables and methods each class should have, how to use inheritance, etc.

If you’re not familiar with the math behind deep learning, Then you should check out this article I wrote earlier (https://medium.com/@amandalmia18/guide-for-deep-learning-aspirants-with-focus-on-non-computer-science-st Udents -87b1f7b3f4b9), which has some related resources.

Otherwise, I feel like watching “deep learning” (http://www.deeplearningbook.org), chapter 2, 3 and 4 of this book is enough to cope with theoretical problem in the interview. I prepared summaries of several chapters to try to explain some concepts that were initially difficult for me to understand, but you can refer to these summaries if you don’t want to read the whole chapter.

If you’ve finished probability theory, you should be able to answer some math questions as well. As for the statistical issues, covering these topics (http://qr.ae/TUTV9f) should suffice.

Questions related to machine learning may depend on the type of job you are applying for. If it’s a traditional machine learning-based interview, they’ll look at machine learning basics. To prepare for the interview, you can complete any of the following courses:

• Wu En of machine learning – CS 229 (http://cs229.stanford.edu)

• The California institute of technology professor Yaser Abu – Mostafa machine learning course (https://work.caltech.edu/telecourse.html)

Important topics include: supervised learning (classification, regression, support vector machines, decision trees, random forests, logistic regression, multilevel perceptrons, parameter estimation, Bayesian decision rules), unsupervised learning (K-means clustering, Gaussian mixture models), and dimensionality reduction (PCA).

If you’re applying for a more senior role, there’s a good chance you’ll be asked about deep learning. In this case, you should be very familiar with either convolutional neural networks (CNN) or recursive neural networks (RNN) and their variants. And you need to know what the basic ideas behind deep learning are, how CNN/RNN works, what architectures are out there, and what are the motivations behind these architectural changes.

There are no shortcuts, either you already know them or you’ve spent enough time learning about them. For CNN, the recommended resources are CS 231N and CS 224N at Stanford. I find Hugo Larochelle neural network courses (https://www.youtube.com/watch?v=SGZ6BttHMPw&list=PL6Xpj9I5qXYEcOhn7TqghAJ6NAPrNmUBH) is also very instructive.

英文原文 :

https://blog.usejournal.com/what-i-learned-from-interviewing-at-multiple-ai-companies-and-start-ups-a9620415e4cc