Let\u2019s get into it!<\/p>\n I\u2019ve said it before, but a machine learning engineer is not exactly an entry-level position.<\/p>\n This is because you need skills in so many areas:<\/p>\n You certainly don\u2019t need to be an expert in all of them, but you should have solid knowledge.<\/p>\n Machine learning engineers are probably the highest-paid tech job nowadays. According to levelsfyi<\/a>, the average salaries in the UK are:<\/p>\n Levelsfyi is generally on the higher end as the companies on their website are often large tech companies, which typically pay higher salaries.<\/em><\/p>\n<\/blockquote>\n With all this in mind, that\u2019s not to say you can\u2019t land a machine learning engineer job right out of university or college; it\u2019s just very rare, and I have hardly seen it.<\/p>\n If you have the right background, such as a master\u2019s or PhD in CS or maths that\u2019s focussed on AI\/ML, you are much more likely to get a general machine learning role, but not necessary a machine learning engineering one.<\/p>\n So, for the majority of people, I recommend you become a data scientist or software engineer first for a few years and then look to become a machine learning engineer.<\/p>\n This is precisely what I did.\u00a0<\/em><\/p>\n I was a data scientist for 3.5 years and then transitioned to a machine learning engineer, and this path is quite common among machine learning engineers at my current company.<\/p>\n Whether you become a data scientist or software engineer is up to you and your background and skill set.<\/p>\n So, decide which role is best for you and then try to land a job in that field.<\/p>\n There are so many software engineer and data scientist roadmaps on the internet; I am sure you can find one easily that suits your way of learning.<\/p>\n I have a few Data Science<\/a> ones that you can check out below.<\/p>\n If I Started Learning Data Science in 2025, I\u2019d Do This<\/strong> How I\u2019d Become a Data Scientist (If I Had to Start Over)<\/strong> Once you have a job as a data scientist or software engineer, your goal should be to develop and work on machine learning projects that go to production.<\/p>\n If a machine learning department or project exists at your current company, the best approach is to work on these.<\/p>\n For example, a friend of mine, Arman Khondker<\/a>, who runs the newsletter \u201cthe ai engineer\u201d<\/a> that I highly recommend you check, transitioned from being a software engineer at TikTok to working at Microsoft AI as an engineer.<\/p>\n According to his newsletter<\/a>:<\/p>\n At TikTok, I worked on TikTok Shop<\/strong>, where I collaborated closely with the Algorithm Team<\/strong>\u200a\u2014\u200aincluding ML engineers and data scientists working on the FYP (For You Page) recommendation engine.<\/em><\/p>\n<\/blockquote>\n This experience ultimately helped me transition into AI full-time at Microsoft<\/strong>.<\/em><\/p>\n<\/blockquote>\n However, for me, it was the other way around.<\/p>\n As a data scientist, you want to work with machine learning engineers and software engineers to understand how things are deployed to production.<\/p>\n At my previous company, I was a data scientist developing machine learning algorithms but wasn\u2019t independently shipping them to production.<\/p>\n So, I asked if I could work on a project where I could research a model and deploy it end to end with little engineering support.<\/p>\n It was hard, but I learned and grew my engineering skills a lot. Eventually, I started shipping my solutions to production easily.<\/p>\n I essentially became a machine learning engineer even though my title was data scientist.<\/p>\n My advice is to speak to your manager, express your interest in developing machine learning knowledge, and ask if you can work on some of these projects.<\/p>\n In most cases, your manager and company will be accommodating, even if it takes a couple of months to assign you to a project.<\/p>\n Even better, if you can move to a team focused on a machine learning product, like recommendations on TikTok shop, then this will expedite your learning as you\u2019ll be constantly discussing machine learning topics.<\/p>\n This relates to the previous point, but as I said earlier, machine learning engineers require an extensive remit of knowledge, so you need to up-skill yourself in the areas you are weaker on.<\/p>\n If you are a data scientist, you are probably weaker in engineering areas like cloud systems, DevOps, and writing production code.<\/p>\n If you are a software engineer, you are probably weaker on the maths, statistics and machine learning knowledge.<\/p>\n You want to find the areas you need to improve and focus on.\u00a0<\/p>\n As we discussed earlier, the best way is to tie it into your day job, but if this is not possible or you want to expedite your knowledge, then you will need to study in your spare time.<\/p>\n I know some people may not like that, but you are going to need to put in the extra hours outside of work if you want to get a job in the highest paying tech job!<\/p>\n I did this by writing blogs on software engineering concepts, studying data structures and algorithms, and improving my writing of production code all in my spare time.<\/a><\/p>\n One thing that really helped me was to develop a specialism within machine learning.<\/p>\n I was a data scientist specialising in time series forecasting and optimisation problems, and I landed a machine learning engineer role that specialises in optimisation and classical machine learning.<\/p>\n One of the main reasons I got my machine learning engineer role was that I had a deeper understanding of optimisation than the average machine learning person; that was my edge.<\/p>\n Machine learning engineer roles are generally aligned to a specialism, so knowing one or a couple of areas very well will significantly boost your chances.\u00a0<\/p>\n In Arman\u2019s case, he knew recommendation systems pretty well and also how to deploy them end-to-end at scale; he even said this himself in his newsletter<\/a>:<\/p>\n This work gave me firsthand experience with:<\/em><\/p>\n \u2013 Large-scale recommendation systems<\/strong><\/em><\/p>\n \u2013 AI-driven ranking and personalization<\/strong><\/em><\/p>\n \u2013 End-to-end ML deployment pipelines<\/strong><\/em><\/p>\n<\/blockquote>\n So, I recommend working in a team that focuses on a particular machine learning area, but to be honest, this is often the case in most companies, so you shouldn\u2019t need to think too hard about this.<\/p>\n If you can\u2019t work on machine learning projects at your company, you need to study outside of hours again. I always recommend learning the fundamentals first, but then really think of the areas you want to explore and learn deeepr.<\/p>\n Below is an exhaustive list of machine learning specialisms for some inspiration:<\/p>\n I usually recommend knowing 2 to 3 in decent depth, but narrowing it down to one is fine if you want to transition soon. However, see if sufficient demand exists for that skill set.<\/p>\n After you become a machine learning engineer, you can develop more specialisms over time.<\/p>\n I also recommend you check out a whole article on how to specialise in machine learning.<\/p>\n How To Specialize In Data Science \/ Machine Learning<\/a><\/strong> In tech companies, it is often stated that to get promoted, you should have been operating at the above level for 3\u20136 months.<\/p>\n The same is true if you want to be a machine learning engineer.<\/p>\n If you are a data scientist or software engineer, you should try as hard as possible to become and work like a machine learning engineer at your current company.<\/p>\n Who knows, they may even change your title and offer you the machine learning engineer job at your current workplace! (I have heard this happen.)<\/p>\n What I am really getting at here is the identity switch. You want to think and act like a machine learning engineer.<\/p>\n This mindset will help you learn more and better frame yourself for machine learning interviews.<\/p>\n You will have that confidence and an array of demonstrable projects that generate impact.<\/p>\n You can always say, \u201cI am basically a machine learning engineer at my current company.\u201d<\/p>\n I did this, and the rest is history, as they say.<\/p>\n Join my free newsletter, Dishing the Data<\/em>, where I share weekly tips, insights, and advice from my experience as a practicing machine learning engineer. Plus, as a subscriber, you\u2019ll get my FREE Data Science Resume Template!<\/strong><\/p>\n Dishing The Data | Egor Howell | Substack<\/strong>First become a data scientist or software\u00a0engineer<\/h2>\n
\n
\n
\n
How I would make my data science learning more effective<\/em><\/a><\/p>\n
<\/a>Roadmap and tips on how to land a job in data science<\/a><\/em><\/p>\nWork on machine learning\u00a0projects<\/h2>\n
\n
\n
Up-skill in opposite\u00a0skillset<\/h3>\n
Develop a speciality in machine\u00a0learning<\/h2>\n
\n
\n
<\/a>Is it better to be a generalist or specialist<\/em>?<\/a><\/p>\nStart operating as a machine learning\u00a0engineer<\/h2>\n
Another thing!<\/h3>\n
Advice and learnings on data science, tech and entrepreneurship. Click to read Dishing The Data, by Egor Howell, a\u2026<\/em>newsletter.egorhowell.com<\/a><\/a><\/p>\nConnect with\u00a0me<\/h3>\n
![\"\ud83d\udc49\"](\"https:\/\/i0.wp.com\/s.w.org\/images\/core\/emoji\/15.0.3\/72x72\/1f449.png?ssl=1\")