I was literally confused about these roles before I even watched this video. Thank you for making this video and helping me get clarity on each role and their specific responsibilities.
A lot of folks doing ML research in universities do not have a basic understanding of what ML engineering looks like in the real world. They should watch this and learn about deep learning systems.
@@TheEsotericProgrammeri’m a regular SWE not ML, so idk for sure, but I don’t think what he said necessarily implies that. To me, it seems like the things MLEs are training just aren’t particularly novel in comparison to what research scientists and engineers do.
@@TheEsotericProgrammer Some do. Some don't. MLE is a really broad term, that's the annoying part. I have spoken to a few MLE's now, and it varies so much haha But in general, an MLE is a SWE with focus on somehow dealing with ML.
Very helpful video, helped clear up the chaos in the head when it comes to job titles. Sometimes it feels like there’s too many. Just a language suggestion though: maybe better to use “they” when referring to ML engineer rather than “he”. Due to understandable reasons
This is the best video I've found on the topic! Can definitely confirm that the details are pretty accurate based on my experience in DE, DS and MLE. Can you make a video on applied scientist role, eg how to develop problem solving skills for such a role, example projects to build such as kaggle competitions, etc? Thanks!
As an ML Engineer, I feel like we're a slightly worse version of each of the other roles, which can actually be really valuable for a company. I've done ML Ops (deploying LLMs into prod), Data Engineering (ETL pipelines, data lakehousing), Data Science (building bespoke models from scratch), Software Engineering (building the system to run everything on) all under the same job role with the same company.
Machine Learning Engineering might not be what you expect; it's more than just coding algorithms. It involves building and optimizing end-to-end systems, managing data pipelines, and ensuring models integrate smoothly into production environments. The role blends technical expertise with practical problem-solving, making it a dynamic and multifaceted career in the tech industry.
Hi Boris! Thanks for the great content, as always. I have always had confusion regarding the differences between various roles in ML and AI; your video clears up many of these. I wanted to know your opinion about something related to this topic. How much does one's PhD major matter in the case of getting a job as a research scientist and/or applied scientist? I am very eager to know this because, at least for now, my dream is to work as an applied scientist or research scientist in the future. However, I am probably going to start my PhD in the College of Pharmacy in the coming fall. My PhD area will be, in general, in the field of healthcare AI for pharmaceutical outcomes and policy determination. I am confused regarding whether I should accept the position or not. Or should I try again next year for a PhD since I don't have any more offers in the ML/AI field this year? Competition in the ML/AI field is increasing tremendously! Feeling hopeless from time to time, to be honest!
Hey :) Thank you! I'm happy my video could be somewhat useful 😊 And regarding your question. I would say, if you want to work on ML for pharmacy, the role you are offered sounds good! You don't necessarily have to be at a CS department to study ML. You will have great domain expertise surrounding you. You would now ideally want to perhaps find some other colleagues that are more experienced with the ML side. But overall, I think it sounds promising 😊 Enjoy it! 💛
i saw your video and i am lost i want to build my own SAAS that provide an AI service, which path should i take to gain necessary knowledge i am a software engineer i don't want to be stuck at a job or doing research my whole life
Hey Boris, I hope to find some help here. I arrived in Australia to start my master thesis. It is about optimizing prices parameters in a manufacturing environment. Where I have some machines for laying up carbon fiber (Multilayer-2D) and another machine with curing/press process, which adds resin to the fiber in order to harden it. Basically I gotta somehow check out the input data and see the correlation. How can I do it? What would you do? The quality is not really measured numerically or binary. I just have some weird data sheets of the machines. I feel overwhelmed. Is there any machine learning algorithm you would advise? The aim is to actually see correlation, check out the critical parameters for the quality, then do some experiments with setting these critical parameters up, in order to validate the analytics part.
thank you Boris, this differentiation between the different roles have been very interesting and insightful, I definitely understand them better. I have a question tho if I may, you said that it's easier for a software engineer to transition into ML. However, talking about my case, I'm usually better at maths, and I'm just starting to learn python, sql and the relevant languages for the domain of data science. For someone who wants to work as an ML engineer, do I need years of experience as a software engineer first? or is it possible to break directly into the field with enough personal projects for example? I know this is a difficult and vast question, but if you have some tips that would be awesome.
Hello, could you help me with a question I have. I know that there are master's degrees that are 100 focused on automatic learning, but do you think that a master's degree in pure mathematics along with small complementary courses could help me ?
In theory, they are not exactly the same. ML is a subset of AI. AI can also include rule and logic-based handcrafted algorithms that don't leverage any learning from data whereas ML refers to artificial intelligence that learns patterns from data "by itself" I hope this somewhat makes sense :)
Thank you so much for the explanation! One small thing though: All your example mini videos depict male engineers/researchers and you also always speak about "he" as in a definitely male person. I know that coming from German you are used to a generic masculine (even though even in German this is highly debatable), but in English it's way easier to generalize. Maybe it would allow your viewers to have a less biased image in their head when they think about an ML scientist etc.
Very good video, and your last points about it being easy for software engineers to move into ML. Anyone who has spent years building systems in many industries will be able to generalize the new data from ML. Researchers and developers need each other, and I rarely find someone who is great at both.
ive been doing ML since discovering genetic algorithms about 10 years ago. It will take at least 3 years to get a good grasp on the details and to build enough "chops" to be able to write training loops and utils for many different models. @@hudyakovnick
To be a machine learning specialist, do I need to be skilled at building machine learning models at a low level from scratch (not using existing machine learning packages)?
I have a genuine question. There is Data Engineer, ML Engineer and AI Engineer. Do you see in future any of these merging together? Should we try to be an ML Engineer or directly jump to being an AI engineer?
Guys, if you do ML, and you are surprised by a TH-cam video. That means that you have research culture whatsoever. Your degree is worthless if you can’t read the literature.
Doing a masters in AI to transition into tech. I’m gonna watch this fully later but it seems like it might be better for me to simply to being a software dev instead of a specialized role like ML engineer or research engineer? Let's say I rack up experiences in data engineering, would I still be eligible for a career path like a dev?
but does a ml engineer also source and clean data and builds the model and then deploys it? Its confusing because sometimes the ml engineer is described as the one who should only deploy the model and the data scientist builds the model. and sometimes the ml engineer works also with data and does the complete machine learning pipeline. and the data scientist does not only build a ml model. so confusing.
"the goal of a data scientist is to generate business insights". not really. ive worked as a data scientist on pure research problems, as well as on production-level code for shipping out solutions.
Hi. I have seen a lot of your videos lately and they are very helpful. Do you think a Nvidia GPU is necessary for a beginner in ML if I plan to train my models on cloud with Google Colab? Also Should I prefer Intel or AMD? Thx
2:12 I’m completely confused. Did three or four takes and retakes of sections. Don’t know what he’s talking about? All mumbo jumbo like. Stopped watching. Won’t be coming back. Sorry.
Depends on where you want to get to. Data Science and ML Engineering e.g. are not quite the same. For MLE I would rather recommend CS. But, in the end, both programs would work for either job if you work on respective projects that fit either role on your own :)
Hey Boris! Love the vid. As someone who doing their undergrad and was confused with the difference between these roles, really appreciate the clear explanation! Also would love to know how your path looked like getting that applied scientist role! Definitely would help me recognise what to do, since I’m not in a very prestigious university but have been doing most of my learning on yt or coursera😅😅
Doing a detailed roadmap is always very tricky because there is no one correct path but I'll see what I can do! Perhaps my video on how I would learn ML if I could start over can help you out for now :)
@@notyourbusiness2687 no. You become an MLE after gaining experience in other areas. There are many experienced MLEs with looking for work icons on linked in so you need to get experience first and then move into that role.
Become better at machine learning in 5 min/ week 👉🏻 borismeinardus.substack.com/
I was literally confused about these roles before I even watched this video. Thank you for making this video and helping me get clarity on each role and their specific responsibilities.
Really happy that the video could help a bit ☺️✌🏼
I've been binge watching all your videos. Learned a lot tbh. Thank you for clearing these concepts so easily.
Thank you so much! I'm really glad you like them! 🥳
i watch his videos and end up with the thought , 'bruh , i still dont know shit'
haha yeah, the current ML world is so new and rapidly changing, it‘s a mess 🥲
Fr bro
Similar response here. It’s a brave, honest starting point to admit this.
The people handing out the jobs don't know either so it's fine
@@semkjaer3581 lol , i hope that that is true
Pretty clear video, well done. Additional roles like MLOps Engineer and AI Engineer are also emerging
Thank you!
Yeah, MLOps is a big one as well!
A lot of folks doing ML research in universities do not have a basic understanding of what ML engineering looks like in the real world. They should watch this and learn about deep learning systems.
Thanks for the explanations man, i was indeed confused about these roles
When a dude has this accent you know he is good at ML
Hhahahahaa
This clears a lot of things up. Thank you for taking the time to make this!
Glad it was helpful!
@@borismeinardus So MLE don't even train their own models?
@@TheEsotericProgrammeri’m a regular SWE not ML, so idk for sure, but I don’t think what he said necessarily implies that. To me, it seems like the things MLEs are training just aren’t particularly novel in comparison to what research scientists and engineers do.
@@TheEsotericProgrammer Some do. Some don't. MLE is a really broad term, that's the annoying part. I have spoken to a few MLE's now, and it varies so much haha
But in general, an MLE is a SWE with focus on somehow dealing with ML.
Very helpful video, helped clear up the chaos in the head when it comes to job titles. Sometimes it feels like there’s too many.
Just a language suggestion though: maybe better to use “they” when referring to ML engineer rather than “he”. Due to understandable reasons
AMMMMAAAAAAZING EXPLANATION !!!!!!!!!!!!!!!!!!!!!! TRULY AMAZING!
🤩🤩🤩🤩🤩
This was a much needed video.Thanks!
So glad it was helpful!
This is the best video I've found on the topic! Can definitely confirm that the details are pretty accurate based on my experience in DE, DS and MLE. Can you make a video on applied scientist role, eg how to develop problem solving skills for such a role, example projects to build such as kaggle competitions, etc? Thanks!
Awesome, thank you! Really happy to get some confirmation from fellow ML professionals!
I'll see what I can do on that topic!
As an ML Engineer, I feel like we're a slightly worse version of each of the other roles, which can actually be really valuable for a company. I've done ML Ops (deploying LLMs into prod), Data Engineering (ETL pipelines, data lakehousing), Data Science (building bespoke models from scratch), Software Engineering (building the system to run everything on) all under the same job role with the same company.
much awaited! thanks!
Hope you like it!
now there is a question for now. If MLE job roles are speciliazed form of SWE, What is the definition of job the SWE :D?
Machine Learning Engineering might not be what you expect; it's more than just coding algorithms. It involves building and optimizing end-to-end systems, managing data pipelines, and ensuring models integrate smoothly into production environments. The role blends technical expertise with practical problem-solving, making it a dynamic and multifaceted career in the tech industry.
What a great explanation! Loved it, keep up the great work:)
Thanks! 😃
Will do my best 🫡☺️
Really excellent explanation. Thank you!
Really glad it was helpful!
this video reads like a resume pre prepared for when the ML bubble pops
Hi Boris! Thanks for the great content, as always. I have always had confusion regarding the differences between various roles in ML and AI; your video clears up many of these. I wanted to know your opinion about something related to this topic. How much does one's PhD major matter in the case of getting a job as a research scientist and/or applied scientist? I am very eager to know this because, at least for now, my dream is to work as an applied scientist or research scientist in the future. However, I am probably going to start my PhD in the College of Pharmacy in the coming fall. My PhD area will be, in general, in the field of healthcare AI for pharmaceutical outcomes and policy determination. I am confused regarding whether I should accept the position or not. Or should I try again next year for a PhD since I don't have any more offers in the ML/AI field this year? Competition in the ML/AI field is increasing tremendously! Feeling hopeless from time to time, to be honest!
Hey :)
Thank you! I'm happy my video could be somewhat useful 😊
And regarding your question. I would say, if you want to work on ML for pharmacy, the role you are offered sounds good!
You don't necessarily have to be at a CS department to study ML. You will have great domain expertise surrounding you. You would now ideally want to perhaps find some other colleagues that are more experienced with the ML side. But overall, I think it sounds promising 😊
Enjoy it! 💛
i saw your video and i am lost i want to build my own SAAS that provide an AI service, which path should i take to gain necessary knowledge i am a software engineer i don't want to be stuck at a job or doing research my whole life
I really wish you had a discord group as well😊
Maybe one day! ;)
You communicate your points and ideas very well.
Thank you 😊😊
This is actually great.! Keep it going
Hey Boris, I hope to find some help here. I arrived in Australia to start my master thesis. It is about optimizing prices parameters in a manufacturing environment. Where I have some machines for laying up carbon fiber (Multilayer-2D) and another machine with curing/press process, which adds resin to the fiber in order to harden it.
Basically I gotta somehow check out the input data and see the correlation. How can I do it? What would you do? The quality is not really measured numerically or binary. I just have some weird data sheets of the machines. I feel overwhelmed. Is there any machine learning algorithm you would advise?
The aim is to actually see correlation, check out the critical parameters for the quality, then do some experiments with setting these critical parameters up, in order to validate the analytics part.
thank you Boris, this differentiation between the different roles have been very interesting and insightful, I definitely understand them better.
I have a question tho if I may, you said that it's easier for a software engineer to transition into ML. However, talking about my case, I'm usually better at maths, and I'm just starting to learn python, sql and the relevant languages for the domain of data science. For someone who wants to work as an ML engineer, do I need years of experience as a software engineer first? or is it possible to break directly into the field with enough personal projects for example? I know this is a difficult and vast question, but if you have some tips that would be awesome.
Really useful info, well presented, thanks
Bro please make a video about ML mathematic 🥲😭.Love from Bangladesh ✨
What are good sources for research papers? Are there free resources or do most need a subscription?
Pretty much all ML papers are free to access! Just enter some topic + "research paper" or just "paper".
You will most likely find them on arxiv.org :)
Simply awesome content! Thank you!
Glad it was helpful!
Amazing video, thank you!
Hey Boris, please where can I find your paper on road network and predicting traffic flow? I am doing research on something similar, it might help 🙏🏻
Bro i want to see how you edit your videos, they are really cool
This is very useful, thank you!
Glad it was helpful! 💪🏻
Hello, could you help me with a question I have. I know that there are master's degrees that are 100 focused on automatic learning, but do you think that a master's degree in pure mathematics along with small complementary courses could help me ?
Extremely helpful thank you
Thanks, very useful!
Stabil Boris!
😤
Vro Is AI and ML are same or different if yes then explain me in short
In theory, they are not exactly the same. ML is a subset of AI. AI can also include rule and logic-based handcrafted algorithms that don't leverage any learning from data whereas ML refers to artificial intelligence that learns patterns from data "by itself"
I hope this somewhat makes sense :)
We are waiting you to launch your machine learning bootcamp and we will pay.
Pretty much my job... all of them 😂
Thanks!
Thank you even more!!
Thank you so much for the explanation! One small thing though: All your example mini videos depict male engineers/researchers and you also always speak about "he" as in a definitely male person. I know that coming from German you are used to a generic masculine (even though even in German this is highly debatable), but in English it's way easier to generalize. Maybe it would allow your viewers to have a less biased image in their head when they think about an ML scientist etc.
🤖🧠
🤖🤖🤖
This kid doesnt know squat
I am sure he can squat just fine.
Faker science
What do you mean?
Awesome video! You did a great job.
Thank you so much! 😊
Very good video, and your last points about it being easy for software engineers to move into ML. Anyone who has spent years building systems in many industries will be able to generalize the new data from ML.
Researchers and developers need each other, and I rarely find someone who is great at both.
Thank you!
Whether it's easy of not depends on the person, I guess haha
But in general, yes! You are spot on 😊
I'm curious, how much time did you spend on this to understand at this level?
ive been doing ML since discovering genetic algorithms about 10 years ago.
It will take at least 3 years to get a good grasp on the details and to build enough "chops" to be able to write training loops and utils for many different models. @@hudyakovnick
Ok so basically ML engineer deploys ML algorithms on applications for users based on the custome'r's data in businesses ,I get it😂😂
By far the best video on differences between various ML profiles....kudos!!!
Wow, thanks! 🤩
To be a machine learning specialist, do I need to be skilled at building machine learning models at a low level from scratch (not using existing machine learning packages)?
In general, no :)
I have a genuine question. There is Data Engineer, ML Engineer and AI Engineer. Do you see in future any of these merging together? Should we try to be an ML Engineer or directly jump to being an AI engineer?
Same question bro
Hey man, love your videos and explanations. :)
Hey :) Really glad to hear it! 😊😊
Thank you for the clear explanation! This is very helpful to me !
What a relief that the video was at least somewhat helpful haha 😬☺️
Guys, if you do ML, and you are surprised by a TH-cam video. That means that you have research culture whatsoever. Your degree is worthless if you can’t read the literature.
ML Engineering: DataOps, back-end, etc.
Data Engineering: Database, data extraction, feature extraction, storage, etc.
Doing a masters in AI to transition into tech. I’m gonna watch this fully later but it seems like it might be better for me to simply to being a software dev instead of a specialized role like ML engineer or research engineer?
Let's say I rack up experiences in data engineering, would I still be eligible for a career path like a dev?
but does a ml engineer also source and clean data and builds the model and then deploys it?
Its confusing because sometimes the ml engineer is described as the one who should only deploy the model and the data scientist builds the model.
and sometimes the ml engineer works also with data and does the complete machine learning pipeline. and the data scientist does not only build a ml model. so confusing.
"the goal of a data scientist is to generate business insights". not really. ive worked as a data scientist on pure research problems, as well as on production-level code for shipping out solutions.
Hi. I have seen a lot of your videos lately and they are very helpful.
Do you think a Nvidia GPU is necessary for a beginner in ML if I plan to train my models on cloud with Google Colab?
Also Should I prefer Intel or AMD?
Thx
Amazing video!
Thanks 🥰
amazing video brother
Thanks 💛💛
Very well put. I usually have to explain this myself but now i have a content to point other people to 😂
So, if i am recruited as a ml engineer in a company, will it be 100% a coding or like programmer role?
Which is the highest paying role amongst this? Is it research scientist in big tech/startups?
Loved your video sir , sir in future video we want you to teach us about these things that you have told us
Please do a video on how much math you use as machine learning researcher
So in addition to ml/dl skills & knowledge we must also know data engineering?
Really informative descriptions👍
2:12
I’m completely confused.
Did three or four takes and retakes of sections.
Don’t know what he’s talking about?
All mumbo jumbo like.
Stopped watching.
Won’t be coming back.
Sorry.
Was really clear, even after 1 time listening.
Should I study a pure computer science degree or get a specialization in Data science for my degree?
I ask myself
Depends on where you want to get to. Data Science and ML Engineering e.g. are not quite the same. For MLE I would rather recommend CS. But, in the end, both programs would work for either job if you work on respective projects that fit either role on your own :)
Ive been thinking of going the data engineering route once I graduate. Is that possible with a BA/BS in CS?
Yes! 😊🤞🏼
Hey Boris! Love the vid. As someone who doing their undergrad and was confused with the difference between these roles, really appreciate the clear explanation!
Also would love to know how your path looked like getting that applied scientist role! Definitely would help me recognise what to do, since I’m not in a very prestigious university but have been doing most of my learning on yt or coursera😅😅
Glad you enjoyed it!
I'll see what video I can make that can help with your questions! Stay tuned 😊
It would be great if you could make a detailed roadmap vid on how to become an applied scientist and a research engineer
Doing a detailed roadmap is always very tricky because there is no one correct path but I'll see what I can do!
Perhaps my video on how I would learn ML if I could start over can help you out for now :)
No ones talking how bros nick is so bigg and looks like f1 driver😂
Maybe I am one 😤
but yeah, no, it does indeed look a bit weird haha
Your my best motivation to learn ML thanks for uploading videos ❤❤.
Happy to hear that! Keep it up! 😊💛
I’m a recruiter in this space and this gentleman knows his stuff. Great video
Can students who are fresh out of college apply to ml jobs?
@@notyourbusiness2687 no. You become an MLE after gaining experience in other areas. There are many experienced MLEs with looking for work icons on linked in so you need to get experience first and then move into that role.
Damn according to this video I’m far knowledgeable then I give myself credit for
Share some projects you built with that knowledge
Yeah! The tricky part is to prove those skills to others e.g. via projects haha
Ikr I was thinking the same thing😂
I feel like data preprocessing and analysing will soon be done by models instead of humans writing algorithms.
I think to which degree depends on the data, but yes, I agree that models with assist us/ take over certain tasks more and more :)