Break Into AI: A Q&A with Andrew Ng on Building a Career in Machine Learning

Title: Break Into AI: A Q&A with Andrew Ng on Building a Career in Machine Learning
Date: Tuesday, December 4, 2018
Time: 2:00 PM ET
Duration: 1 hour

SPEAKER: Andrew Ng


1 Like

Question for the “Break Into AI” Webinar: What does it really take for me to develop my own AI/ML solutions for analytics and operations for network infrastructure solutions, which is my area of work? AI/ML solutions requires data cleaning and multiple steps to create optimal sequence of AI/ML algorithms to create valuable solutions for real world networking solutions.I am taking Master of Science in Data Science at Illinois University, where I will focus on Machine Learning, both applied ML and the “hard” matematics behind in the format of advanced Bayesian Modelling. Which other skills do I need to achieve to meet my goal, e.g. in matematics and development techniques (I have taken Your Machine Learning course at Coursera approx. 1 year aogo). BR Jesper

Question for the “Break Into AI” Webinar, and/or the attendees that read this, as this seems to be the way to provide questions ahead of time (low priority):
I don’t believe that anyone here will argue, AI can be applied in a useful way to almost any topic. I am more interested in pulling AI into my field (website/application dev. for a school board) than I am interested in breaking into AI and seeing where that takes me. Before I can even start this though, I need some assistance answering some of the early questions.

a) How can I argue that we stand to benefit from introducing AI - to a management class that might not understand AI beyond Sci Fi movies? Alternatively, what kinds of problems could be solved for a school board by dipping my toes into AI?

b) What would you recommend someone in a non-AI-centric company should do to get started, both in understanding how to apply AI and where to start applying it (possibly suggesting a relatively simple demo topic)?

c) What are some of the most common “starting with AI” questions, struggles, and complications that one should be prepared for when Breaking into AI?

Thank you for your time,


Question: Do writing solutions such as Hemingway and Grammarly showcase AI application? If so, what is the summary of current AI related language products?

Can you suggest where a practicing Sr. Technical Program Manager and Solution Architect that also has significant product and IT software development and leadership experience can best focus in the Machine Learning/AI field - to better understand key concepts/technologies/approaches/frameworks so as to make the right strategic solution architecture and program/product management decisions that better leverage Machine Learning? This can certainly include some hands-on development/experimentation, where it can provide the greatest pedagogical benefit; but I’m interested in leveraging my MBA as much if not more than my MSE in my current role. Thanks - Brian

It seems that the basics of machine learning have rapidly gone from experimental to commoditized, in that there are established platforms that anyone can learn in a few weeks. The course I’m taking now quickly breezed past the math, and got us into building networks with pytorch in the third week.

Question: Given the rapid changes in the field, what machine learning areas / techniques do you see as being valuable or distinguishing to developers 5 years from now?

1 Like

Question: How important is having today a MS or a Ph.D in Computer Science/Math/Machine Learning in order to work as a Machine Learning Engineer?

Question: How difficult is it for a non-programmer/System Administrator to learn/adapt Machine Learning? There are several resources available online and everyone has a different perspective. What is your advice for such students who want to go from novice to pro?

Question: It is quite confusing to see the terms ML and AI often used in same topics. Can you give us a short insight about the key differences between ML and AI?

Question: ML or AI demands interdisciplinary knowledge such as statistics, mathematics, data science, computer science, et al. From which discipline should I start to learn at first while trying to build a career in ML as a beginner?

Career question: With a Master degree in Physics with machine learning, what would you recommend for best chances of interesting and secure work opportunities the next 40 (!) years:

  1. some years in finance (sales & trading) for cross-diciplinary experience, or
  2. phd in physics and/ or machine learning as soon as possible

Question for the “Break Into AI” as well. I am a web developer who has some academic experience with ML and python, wondering what roles or future roles would be a good fit for a software engineer with data science experience? Is it advantageous for someone who already has good experience as a JS developer to move to a path that can combine data science and programming to become a software engineer in the field of AI? What sort of qualifications and experience are best. Would a graduate degree in data science help or would that be becoming too much of a ‘jack of all trades’ with little expertise. How does someone who has programming experience best prepare themselves to transition to a job that is just as good in AI without starting again as a Junior?

Question for the “Break Into AI”
I am a coding student from morocco i study at 1337 (
I have some ML/DL knowledge but since i study at a new and have mandatory 6 months internship, my question is how can i make my self more attractive to companies at silicon valley to get that internship.

i am fluent at python, c and js , and i have some interesting side projects. Would that be enough ??
Thanks in advance mr. Ng

Question: Can you talk about the Ethical Artificial Intelligence? highlighting the following points:

  • your say on ethical implications of AI and fears and doubts surrounding the field
  • application areas of AI that you feel more vulnerable to ethical issues
  • measures to be taken to adopt AI in a socially responsible way

Question: Hi there. I am currently interested in doing a Masters Thesis Research Project in utilizing Artificial Intelligence to diagnose Cancer. Thus far, for the past year, I’ve been trying to learn Python and currently, looking into learning the various Machine Learning algorithms by use of the different Python Machine Learning libraries. I’ve also watched and read through some materials on Cancer. I’m thinking of focusing my research specifically on Breast Cancer in this regard, and have found the following dataset that I could use:

I am a Computer Science professional and have a degree in Computer Science as well as an Honours degree (Cum Laude) in IT.
My research project above is for my Masters Thesis.

My supervisor has advised on using an existing Data Model - a Data Model that has already been published. And he has suggested that I focus my Masters Thesis Research Project on improving such a Data Model.

May you please kindly assist me in regards to the above - i.e. what should be my approach to trying to learn Machine Learning and then pursuing this Project - i.e. improving the Data Model?
Also, how and where do I access the Data Model that has been already published by other researchers?

Where would you suggest I get the best Machine Learning in Breast Cancer scientific research papers from?

Please assist.

Many Thanks in advance. It is greatly appreciated!

P.S. Depending on my journey in doing my Masters project, as mentioned above, I will also look to see whether building a Career in Machine Learning interests me as a whole as well.

Thank you so much in advance. Much and truly appreciated.

Question for the “Break Into AI” Webinar:
Dear Prof. Andrew, I have the questions below:

  1. What is the best way to get myself more familiar of Tensorflow and other frameworks (familiar means more than being very comfortable with it and can use it very proficiently with people’s own code, etc.)? Should we just focus on one framework at each time?
  2. Also, given the current trend, Python is becoming much more important/useful than Java in Machine Learning field, any more suggestion on must-have skills?
  3. For people who just started career in ML (like just started their internship or full-time job), what should they do on a regular basis to keep up with this fast developing trend?
    Thank you!

Question for Dr. Ng – Can you share the name of the drawing tool you are using? Thanks!

I enjoyed the Webinar-- Andrew Ng has a lot of good advice, delivered very crisply.

OTOH, I was puzzled by the moderator’s approach. It did not seem he was passing on questions from the audience, but rather posing his own (often at some length.) Maybe he was trying to paraphrase or clarify audience questions, but maybe it would be better to just pass them on as asked.

I posted two questions at the very beginning of the Webinar via the Q&A link provided, but AFAICT neither was addressed. I realize that there may not be time to address all audience questions, and some may be deemed out-of-scope or otherwise filtered, but I was left wondering if the channel through which I submitted my questions was the right one and if it was being monitored.

In any event, a very worthwhile hour–thanks!


Thank you for the feedback, and we will keep it in mind in the future! I’m happy to be transparent about my process.

We had 500+ questions to go through in real-time throughout the hour, and there were some questions that we had to consolidate and some that I could not get to due to the volume. The questions were of varying length, so it can be a challenge to parse out the key takeaways while maintaining a conversation flow.

I genuinely appreciate your perspective! Thank you for sharing. I hope to see you at the next Webinar, and that you will continue to stay engaged with our discussions.

Juan-- thanks for the detailed response!

It seems the consolidation in real time is what I was noticing.

As I said before, a very worthwhile hour for me.


Merhaba makine öğrenmede bizler sistemleri ve yayınları takip ederek öğreniyoruz bu şekilde makine öğrenmek biraz zor görsel olarak bir çalışma ile insanın aklında daha kalıcı olacagina inanıyorum