Data Scientist vs. Machine Learning Engineer

DeepLearning.AI and FourthBrain invite you to join us at the live event focused on comparing Data Scientists vs. Machine Learning Engineers, including career progressions and career opportunities.

The speakers plan to discuss:
- What are the main differences between the two roles?
- What each role does in a typical workday?
- What is the typical career progression of each role?
- What are the skills required for each role?
- What roadmap do you recommend for someone transitioning from DS to MLE?
- How can a non CS or non engineering background person with no prior experience of the DS or ML start?
- With the lines blurring between the roles of a Data Scientist and MLE, what's the future trends of these two roles?

Speakers:

[Deeplearningai]

7:15 Can each of you use one sentence to describe what Data Scientist and Machine Learning Engineers is in your own words.
9:05 why do you think the concept or description of these two roles is still vague in the community?
10:58 Can you give us examples of what each role does in a typical workday?
13:14 What is the typical career progression of each role?
15:55 Many of our audience today are thinking about starting a career either in DS or ML. To help them figure out which track is a better fit for them, what are the skills required for each role? How much math and software engineering is required?
20:16 What is the minimal educational qualification that differentiates the two roles?
22:08 Do you agree that compared to MLE, a Data Scientist does not require a skill set specific to MLOps?
24:46 How can a non CS or non engineering background person with no prior experience of the DS or ML start?
28:33 Recently, there are a large number of people with technical backgrounds transitioning to data science related roles. In your experience, what is the most critical missing skill / knowledge of these candidates who consider such transition?
32:25 For someone who's currently a Data Scientist focused on Machine Learning modeling, but wants to explore the potential career path in Machine Learning Engineering, what roadmap or resources would you recommend?
37:34 With the lines blurring between the roles of a Data Scientist and MLE, do you think that one could be both a data scientist and MLE?
39:45 Will data scientist roles get dissolved into software engineer roles in the future?
43:40 WIth model building becoming more automated now with advanced packages, is DS/ML moving to become more focused on business intelligence more than technical involvement? Do we need more specialized ML practitioners (Scientist, Engineer, Data Engineer, Software Developer) or we should have an end2end generalist?
47:35 Which industries do you see the most need for and growth in the use of machine learning engineers and data scientists?
49:45 Do you think that having a PhD and academic experience is an advantage or disadvantage (in the sense that you are overqualified) for an industry job?
53:10 Can you also brief the difference of MLE, and DS with Data engineer, Data Analyst?
54:00 How can I get into these roles without prior experience? I have done quite a few courses, but havent had opportunities to apply it in a professional setting.
56:45 Which of the two career paths allows for more creative freedom?

[Sweet-Vermouth]

Unpopular Opinion - MLE in my opinion is as creative a role as Data Scientist.
The way you deliver an ML solution can vary so much. At this point, there is very little standardization. Every model has unique resource requirements and use-cases. You can choose to deploy at the edge, or on the server-side, or you can do federated learning which is a combination of edge and server-side learning. You can choose how you want to build a deep learning pipeline and how you want to optimize the run time. Having done both data science and engineering, I have to say I enjoy both equally and MLE is as taxing and as exciting as Data Science.

[TheLastWhiteKid]

We need more info on entry level positions and how to get started. So many businesses are expecting PhD's and 3-5 years of employment experience for an ENTRY LEVEL JOB. This is ridiculous.

[DistortedV]

Great questions. I wish a more diverse panel of people at FAANG too, or startup/public sector Data scientists but covered pretty well

[Sweet-Vermouth]

The job descriptions make the whole discussion a lot more complicated. Every company posts a different role description and you end up discussing your role with the interviewer when you finally meet them. The live chat was actually a lot of fun. A little more than the video IMO 😅

[benxneo]

Would love a talk relating to Data Engineering next!

[EdgarGuevaraCodina]

Thank you for your answering my question 49:42
It was a great panel all-around!

[zhaoxiao]

I disagree that data scientist ist more "science" than ML engineers. for ml, the model and data has also to be engineered or analyzed. what is science? if data scientist can not publish good papers, just do modeling and data analysis for different business cases, as usual in companies, I would not call them as a scientist. so the job title of data scientist ist more data engineer, if you do not invent new math formulations like that

[NatashaMugwe]

Informative discussion. Craft your path and learn

[divyanshugupta]

Great to know an industry-based insight into the subject.

[Mzulfreaky]

Thanks for the talk! Looking forward for more of this series of discussions

[avniagarwal]

which 16 week program does Kate strachnyl mention at 36:35

[TheRollsP]

Can a software engineer move to MLE and then to DS? Is that path possible and feasible?

[DistortedV]

I disagree that things would become more and more automated soon. Why? Deep learning is not like making websites where more and more pieces can be automated and details forgotten about. You are giving these built in models data, and the data + model synthesis is generating the program. You don't know if it's going to be a good or bad program apriori because deep learning theory is still not well understood. Until then, these automl packages from google or huggingface are kind of like cargo cults to feed to business people.

[sandipansarkar]

Great session. Enjoyed every bit of it

[ashishgoyal]

Summary of the Live Session:
"YES OR NO"