Software Engineering or Machine Learning: What's a better career?

hi guys for those of you who don't know me i'm carolina and i work with data many of you have asked me what is a better career choice machine learning or software engineering when i was doing my master's degree in computer science i had the exact same dilemma and many of my course mates did as well both careers are lucrative both are rewarding yet they are completely different now i've been working in the industry for a little bit now so hopefully i can bring some useful insights to the table so in this video we will compare machine learning and software engineering careers taking into account the following factors one salary and growth in job openings two barriers to entry three your predispositions four predictions for the next 10 years if you're ready then let's get started money is important and many people are drawn into technology precisely because of money and there's probably nothing wrong in that so let's just have a look at the salary breakdown of machine learning engineers versus software engineers in the us and in the uk so first let's have a look at the american market on average software engineers earn 92 000 per year that can be contrasted with 114 000 dollars for machine learning engineers again on average in the uk the situation looks quite similar the salary for software engineers is on average 43 000 pounds compared to 47 000 pounds for machine learning engineers this data was taken from glassdoor it obviously doesn't take into account the geographic breakdown within each country so for example if you work as a software engineer or machine learning engineer in the silicon valley obviously you are going to earn much more than what i just quoted this is just an average and same for london in case of the uk the point is that as you can see in both countries machine learning engineers earn marginally more than software engineers but what about the demand for machine learning engineers versus software engineers in order to analyze that let's have a look at the number of job openings as quoted by indeed.com so according to that website between 2015 and 2018 the number of job openings for machine learning engineers rose by 344 can you imagine that and the number of software engineering positions rose by 207 in the same period remarkable growth in both areas but again the number of job openings for machine learning engineers grew a bit more so when it comes to the financial aspect it appears that machine learning engineers win the second aspect we are going to consider are your predispositions now i believe that this is the most important factor in the whole decision process and what i mean by that is that machine learning and software engineering are completely different fields they call for completely different skills and you need to have completely different mindsets to solve machine learning problems versus software engineering problems now you might be naturally more suited to do machine learning or maybe you're more naturally suited to do software engineering so let's compare and see what is machine learning well machine learning is statistics period statistics means maths if you are not good at maths forget about machine learning and i'm not saying that you cannot learn it because you probably can if you put enough effort into it what i'm saying is that you'll find it extremely difficult and what's the point so machine learning is for someone who loves complex mathematical riddles thinking in abstract theoretical ways so what is software engineering well software engineering is a type of engineering and that means putting things together so that they work it is a much more practical area what is important is your creativity practicality it is not important that you know maths because usually that's not really involved in the process so that's one distinction between machine learning and software engineering [Music] another distinction is that in software engineering you get this instantaneous feedback from the system so you do something you code up your solution and it either works or it doesn't work basically you get this feedback and and you know that you did something right if you did something right and that makes you happy whereas in machine learning there's so many arbitrary things that happen in in machine learning if you're getting very shady results it might be that you didn't clean your data right or it might be that the data is inherently messy and you can't do much about it or it might be that you've initialized your hyper parameters to wrong values or it might be that this algorithm is not the best fit for the data and it can be you know a series of things and you just don't know and you will never know for sure which one of these it is and because you don't know whether you're doing something right or wrong it doesn't give you that instantaneous satisfaction instantaneous feedback the number of technologies that you have to know to do machine learning is actually quite limited which is good or bad depending on how you look at it but basically all you have to know pretty much is python and sql your job is to come up with algorithms so all the maths is more important and then the idea is that you just have to put it into a code python is sufficient to do that one downside of software engineering is that you have to know a lot of technologies especially if you are a full stack developer which means that you have to know the backend and front end so you have to know sql some object-oriented language like java or c-sharp and maybe python and if you do web technologies as well then you also have to know javascript and it's gazillion of frameworks so you know being a software developer is a never-ending learning curve can a curve be never ending it's just that there is it's a long curve there is no horizon oh oh one second my lump just decided to die i can't blame it it's so hard so whether you're better suited to machine learning or software engineering is dependent on whether you a love complex maths then machine learning is for you or b you love creating stuff building stuff then software engineering is for you overall a draw [Music] the third consideration is barriers to entry for machine learning the barriers to entry are quite high very often in job openings you'll see that what is required is an advanced degree in mathematics statistics computer science or other quantitative degrees sometimes they even explicitly ask for a phd in those areas now that's not always the case and i even made a separate video about how to get a job in machine learning without a specialized degree in those quantitative areas however it is often the case that this degree is at least helpful that can be contrasted with software engineering sometimes the job openings ask for a degree but rarely it is an advanced degree like masters or phd and even if you don't have a bachelor's degree often it is enough to have a portfolio of projects that you've coded and so that you can basically have an evidence that you can code so in this category i believe software engineering wins so let's talk about the predictions for the next 10 years [Music] unsurprisingly machine learning is still on the rise given the crazy amounts of data that we produce every day every day 500 million tweets are sent four petabytes of data are created on facebook 294 billion emails are sent four terabytes of data are created on every connected car 65 billion messages are sent on whatsapp five billion searches are made so we generate a lot of data and someone will have to be drawing insights from the data for governments for businesses and i also think that many machine learning engineers at smaller companies in particular will have to learn the data engineering skills because data engineering skills will be extremely important software engineering isn't going anywhere for the next 10 years as well however the trend is that we will be coding less and less it is happening now and it will be certainly even more prominent in the future so basically what software engineering is going to be about is gluing things together understanding the ecosystem how various applications talk to each other and you know what i think that's fun um i'm actually looking forward to that change because i think it frees up creativity i personally prefer this aspect of building things rather than coming up with the most efficient solution you might be different but well i value efficiency but not to the point when it basically slows you down that it's not efficient anymore like i wouldn't spend days and days trying to come up with this perfect algorithm i just want to move on okay so here at draw again both software engineering and machine learning will be in demand in the future [Music] so we have a tie between the two careers now to pick the right one for you ask yourself these questions do you prefer mathematical riddles or building things do you have a specialized degree or can you afford to get one or do you not because a degree might be a prerequisite for many machine learning jobs it doesn't have to be but it might be answering these questions should help you to decide between the two carriers if you like this video please give it a like and subscribe if you haven't already if you're new to this channel this is very important to me it's a very important data for me because then i know that this is the kind of content that you are interested in watching and i can focus on in the future as always if you have any comments questions problems whatever let me know in the comments and i will see you in the next video bye

Channel: Karolina Sowinska

Views: 72,763

Keywords: software engineering or machine learning, machine learning or software engineering, software engineering career salary, machine learnning career salary, machine learning career us, software engineering careers us, machine learning career uk, software engineering careers uk, se or ml, which career in technology, machine learning engineer or software engineer, software engineer or machine learning engineer, software engineer salary, machine learning engineer salary

Id: FkqF2Nq062o

Length: 12min 49sec (769 seconds)

Published: Thu Aug 13 2020