I must be doing something right if people are getting in touch with me out of the blue with questions, both technical and otherwise, about the Udacity Self-Driving Car Engineer Nanodegree Program. It’s great! I’d already met several interesting people, but when my EKF project blog post kinda maybe blew up a little bit I had even more coming out of the woodwork. This conversation I had with John (not actually John Cena, by the way) was too good not to share:
I got your email off your website.
I’m interested in doing the self-driving car Nanodegree and I saw on Medium you’re doing/completed it.
I was wondering how many hours per week you put into it? As in, would I be able to work full time while doing it?
I’m probably a bad example because, while I work full time, I have plenty of time to watch the lesson material while I’m at work. Also, it really just depends on how much extra effort you want to put in (beyond just the basic requirements) and how well you know some of the fundamentals going into the program - stuff like Python, command line, git/Github, some basic geometry/trigonometry/calculus/statistics/matrix math, and basic machine learning. I had never used Python, and knew nothing of machine learning when I applied, but while waiting for my cohort to begin I took Intro to Machine Learning, Deep Learning, and Artificial Intelligence for Robotics (and started Intro to Computer Vision). That helped me a LOT. I probably would have struggled to keep up, otherwise. Quite a few of the lessons I finished beforehand were already marked completed when I began the Nanodegree.
I would say that I put in an average of 15 hours per week. Sometimes I pushed a little harder, maybe even up to 25 hours some weeks, but I was also trying to stay about a week ahead of the project deadlines. I ended up finishing Term 1 a few weeks early (and was able to start Term 2 early). I think that I probably could have gotten by doing the 10 hours per week Udacity prescribed, but I probably wouldn’t have had time to work the bonus problems and might have been late on an assignment or two.
I think it’s entirely possible to work full time and finish the program (even if you don’t have spare time at your job), especially if you’re motivated and the material is interesting to you. I think a good way to test the pace and your interest is to take the Artificial Intelligence for Robotics course. I plowed through it and finished in two weeks, and that was roughly the pace I kept in the Nanodegree. I think if you can finish that course within a month, you should be fine (plus, you’ll have a nice foundation for Term 2). Intro to Machine Learning was way easier, and probably not the most relevant to the nanodegree (but decent Python practice), and Intro to Computer Vision is really long (almost too long to bother with, but it WAS helpful), Deep Learning is helpful but short and dense (trial by fire, but it’s good for having some familiarity going in), that’s why I’d recommend AI for Robotics - it’s kind of that sweet spot of relevant and appropriate length.
Sorry, I’m rambling and you probably already knew some of this if you’ve read my blog. I hope I answered your questions - feel free to hit me up if you have others.
Thank you so much Jeremy,
I’m incredibly keen to get into the industry but it would be impossible if I can’t support myself. From what you have written it sounds like it is possible. I will potentially have a poor internet connection at times. Is the content downloadable or is it 100% streaming? are you using a mac or pc?
I’m about to finish a Masters of Mechanical Engineering so I’ve got a lot of the basics down but definitely lacking in the programming/machine learning side. I was planning on starting the Nanodegree closer to the end of the year so I will definitely take your advice and complete the courses you recommended. Do you feel like the content you are learning is relevant? How competitive is the course so far?
I hope this question isn’t too forward, but has there been any discussion of employment within the course? or with the sponsoring companies? I’m interested to see if the content taught covers the requirements Self-driving car companies have for their employees.
Thank you again for answering my questions.
You’re welcome John, I’m sure you’ll do great! (Especially if you do some preparation in advance.) Some people have complained about the pace in Term 1, but I get the feeling a few of those people weren’t really putting an earnest effort in and probably weren’t sticking to a consistent schedule. And if you plan to enter the program later in the year, I would bet that by then Udacity will have evened out the load so that it’s not so hectic in Term 1 (Term 2 is already far more relaxed). I’m pretty sure that the vast majority of students in the program work full time.
Yes, as far as I know (I haven’t tried myself) all of the content is available for download and offline viewing. It’s all web-based, and all of the lectures are (private) YouTube videos, but they also have a mobile app you can use to access the material. The projects usually involve setting up an environment on your home machine, but they do a good job of stepping you through the process and make a great effort to be platform agnostic (I’ve used Windows, Mac, and Linux machines with no issues).
I do feel like what we’re learning is relevant to industry. At times it feels like we’re just scratching the surface, and that a real self-driving car would take what we’re learning a billion steps further, but I also get that it’d be impossible to teach us how to do all of that in any finite amount of time. I’m not sure what you mean by “How competitive is the course so far?” Do you mean competitive, as in, in relation to other educational resources? (in which case, Udacity claims to be the only name in this specific game and my feeling based on job postings I’ve seen is that, yes, it covers the vast majority of skills companies are looking for) Or, as in, competitive among the students taking the course? (in which case, I’d say there is a certain subtle element of competition, but for the most part it’s a very, VERY collaborative environment - I think people get that helping others only reinforces your own knowledge, and I even get the sense that we’re all genuinely rooting for each other because, well, this is important work!)
Not too forward a question at all! (I’m an open book.) There’s a lot of discussion about employment, including a Slack channel devoted specifically to careers. Some people (ridiculously, I’d say) seem to expect Udacity to more or less hook us all up with jobs, but Udacity already does a lot to prepare us for finding jobs (that’s the primary goal of the program - to put students in jobs). They offer resume/LinkedIn/GitHub advice and personalized reviews, and they post jobs from hiring partners on their jobs board for which, I assume, students get priority consideration. But not only that, there are a TON of companies out there (more all the time) focusing on self-driving cars (I put together a spreadsheet) and I have no doubt that anybody who’s willing to put in the work will have no problem finding a job. (That said, I also highly recommend keeping a blog throughout the program - I’m really starting to see some interest because of it.)
(Speaking of… unless you object, I might use this conversation in a future blog post. I won’t use your full name, of course.)
Thank you so much Jeremy,
I think you’e answered all my immediate questions. I’m going to digest everything and might send you a few more questions later down the line if thats okay. I expect you’ll be a full Self-Driving Car Engineering at that point!
Yeah please feel free to use our conversations on your blog! I’m certain they will help other keen Engineers!