Data Science Courses/Degress/Work

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
thank you.. I’ve received an interview for bi analyst, is it similar to data analyst? would you recommend me to take it up if my end goal is being a data scientist?
Bi Analyst is actually a very generic role, therefore the actual scope depends on the company. But ideally it should involve some data visualisation and SQL query. I recommend you ask in the interview whether you will have a chance to do this. If so, then I actually think it is a very suitable role for you to gain some experience before you attempt to transition to data scientist.
 

idiotical

Member
Joined
Aug 30, 2007
Messages
308
Reaction score
0
@DataScience

I used to take up some SQL course in datacamp but the env is not very useful when it come to practical. Is there any course that taught directly using postresql or microsoft SQL
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi @DataScience,

I managed to chance upon this lively thread started by you :) really nice of you offering your advice!

Just a few questions here - i am currently in a marketing analytics role as a regional lead, but have secretly habour dreams of doing a full stack data scientist/MLE role for the longest while. My current role is largely about setting up management reporting dashboards and doing timely reporting to C suite management folks, and occasionally i do get to work on some ad hoc ML projects with data scientists like customer segmentation and propensity modeling etc (they do the model deployment obviously hah!). After all these years, i am getting really quite sick with the reporting bit and would love a more hands role with the ML work.

Any advice of transitioning to a more full stack DS/MLE role? i understand for a start, I need to be able to take my pickled model to production level - but there are so many tools out there when it comes to model deployment and stuffs that its quite confusing to a noob like me, and there is so much to learn that it feel daunting at times. I am considering leaving my current role to do a full time bootcamp - almost impossible for me to do part time because i do 70 hours work week (e-commerce co, won't name it LOL)

Any detail knowledge on GA DSI curriculum though? i did asked them whether it covers model deployment/big data etc, but received vague response from them. Another bootcamp I am considering is fourthbrain.ai MLE programme- its based in the states, not sure you heard of it though but I am really liking their curriculum, and its way cheaper than GA - I don't qualify for IMDA TIPP subsidy because they consider me as one who is already in the data industry ~ (roll eyes)

A little bit about my academic background - already have 2 postgrads in applied economics and biz analytics, so definitely won't be doing considering doing another masters haha. Even if i want, my wifey won't allow lol. Apologies for the long post !
Hi,

Not sure why I missed this message. I know you said you won't want to do another masters, but if you want the path with the highest success rate then you should consider doing a full-time OMSCS. This will give you very strong creds for progressing to MLE or even software engineering role. I can tell you with confidence that the strongest MLE out there are those who have spent some time working on Software engineering. I feel that this is the cheapest and safest option for you. Given your work experience + this masters, I am quite confident you will be pretty sought after at the end of your masters.

Your next best choice is probably 4th brain. 4th brain it is a very good school affiliated with Andrew Ng. One of my friend from ivy league went on to study in 4th brain. I would definitely recommend that over GA DSI. At your level DSI won't help you anymore. GA DSI doesn't teach ML Ops. They only teach you how to build research models. They won't really spend much time on the ML Ops. The reason is because their curriculum is rather old, and they don't really change much. And back when they first came up with this curriculum, deployment was hardly the focus. Deployment only gained much attention in the past 3 years.

Haha but If I am in your position, I would try to convince your wife. It really will give you the best outcome. You can go linkedin and scour for some information. Look at the number of people who managed to become DS/MLEs after their OMSCS. Use this information to convince your wife.
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi @DataScience,


Can I ask for your opinion if having a doctorate degree (Phd) will give a big boost in the career in terms of earnings and progression in the AI/ML field? Or is it only applicable to certain sectors?

The reason I am asking is because my boss is asking me on my interest to pursue a Phd while working with them, and currently I am waiting for my OMSCS with Georgia Tech to start this Fall 2022. If I agree to go ahead with the Phd program (provided I can find a suitable mentor who is willing to take me as a part time Phd student), it will mean I may have to drop my coursework with OMSCS since the Phd program will likely take 4 to 5 years to complete. I don't think it makes sense for me to go for OMSCS first then do the Phd program since I may take about 8 years to complete this journey.
What is field is your Phd in though? If it is in a totally unrelated field then it won't help you at all. And it also depends on which school your Phd will be with.
 

p0w3rL3s5

Junior Member
Joined
Jul 24, 2005
Messages
36
Reaction score
0
Hi @DataScience, read through the whole thread and want to thank you for all the sharing. Also hope to get your advice for my case. i'm in my early thirties, and come from a totally irrelevant field without any experience looking to break into data science / AI. I graduated from a local university with a business degree (2nd upp.) and mainly held customer/client facing roles in the finance industry for the last 7 years. Have since left the job and now doing some private tutoring.

After reading through the thread, i have laid out a simplified plan:

1) Start with Heicoders Academy AI100 to AI300
2a) Start building up Kaggle portfolio
2b) Take MOOCs or courses (Datacamp etc.) to fill in knowledge gap
3) Apply for TFIP or AISG Apprenticeship
4) Apply for entry position in industry (after traineeship/apprenticeship)
5) Apply for OMSCS after 2 years in industry

A few qn:

a) Is up to step 4 realistic for a 2 year timeline (assuming i can spare 30-40 hours a week for part 1 and 2)?
b) Is it possible to try and enter the industry after i finish step 2? Or is it better to transit via a structured program like TFIP or AISG Apprenticeship to gain some experience?
c) Can i actually apply for OMSCS earlier or is it better to only apply after at least some years in the industry?
d) Any other advice / recommendations?

Many thanks again!
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
@DataScience

I used to take up some SQL course in datacamp but the env is not very useful when it come to practical. Is there any course that taught directly using postresql or microsoft SQL
Mmm, most courses outside teach theoretical SQL. They don't really teach you how to apply SQL in the industry (how to setup, design and deploy). The best advise i can give you is to work on some SQL project. While you are working on the projects, you can google for best practices on stackoverflow.
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi @DataScience, read through the whole thread and want to thank you for all the sharing. Also hope to get your advice for my case. i'm in my early thirties, and come from a totally irrelevant field without any experience looking to break into data science / AI. I graduated from a local university with a business degree (2nd upp.) and mainly held customer/client facing roles in the finance industry for the last 7 years. Have since left the job and now doing some private tutoring.

After reading through the thread, i have laid out a simplified plan:

1) Start with Heicoders Academy AI100 to AI300
2a) Start building up Kaggle portfolio
2b) Take MOOCs or courses (Datacamp etc.) to fill in knowledge gap
3) Apply for TFIP or AISG Apprenticeship
4) Apply for entry position in industry (after traineeship/apprenticeship)
5) Apply for OMSCS after 2 years in industry

A few qn:

a) Is up to step 4 realistic for a 2 year timeline (assuming i can spare 30-40 hours a week for part 1 and 2)?
b) Is it possible to try and enter the industry after i finish step 2? Or is it better to transit via a structured program like TFIP or AISG Apprenticeship to gain some experience?
c) Can i actually apply for OMSCS earlier or is it better to only apply after at least some years in the industry?
d) Any other advice / recommendations?

Many thanks again!


This is a solid plan you have for yourself. I just have some comments:
- Scratch datacamp, once you are done with AI300 and able to build kaggle projects, datacamp is too basic for you. You would find more useful and relevant courses on coursera.
- I feel that after you are done with AI200, and have built a few kaggle projects, you can try applying for OMSCS already.
- Concurrently, it would be great for you to take a few coursera/udemy fullstack development courses (this will show Gatech you have the passion, and also it will greatly help smoothen your learning curve for OMSCS)
- Given that you achieved 2nd upper in your bachelor degree, it should be quite favourable for your application

This will greatly cut short the runway you need to land yourself in the industry.
 

p0w3rL3s5

Junior Member
Joined
Jul 24, 2005
Messages
36
Reaction score
0
This is a solid plan you have for yourself. I just have some comments:
- Scratch datacamp, once you are done with AI300 and able to build kaggle projects, datacamp is too basic for you. You would find more useful and relevant courses on coursera.
- I feel that after you are done with AI200, and have built a few kaggle projects, you can try applying for OMSCS already.
- Concurrently, it would be great for you to take a few coursera/udemy fullstack development courses (this will show Gatech you have the passion, and also it will greatly help smoothen your learning curve for OMSCS)
- Given that you achieved 2nd upper in your bachelor degree, it should be quite favourable for your application

This will greatly cut short the runway you need to land yourself in the industry.
Hi @DataScience ,

Thanks for the reply! Will defiinitely take into your recommendations. And In case you might have missed the qn regarding trying to enter industry directly after AI200 and Kaggle projs vs entering a structured program first (TFIP, AISGP). Which route do you think would be more suitable?
 

DDMatrix

Junior Member
Joined
May 16, 2017
Messages
77
Reaction score
41
What is field is your Phd in though? If it is in a totally unrelated field then it won't help you at all. And it also depends on which school your Phd will be with.
It will be likely with the Industrial Phd Program under NUS school of Computing. The other option will be under a hybrid healthcare/data science program such as Duke NUS Quantitative Biology and Medicine. I am currently working in the healthcare sector. I am just weighing the pros and cons vs OMSCS which is a shorter 3 year program hence wanted to make an informed decision before taking any action. If completing the OMSCS is enough for a boost in career progression and if the Phd program is likely to give marginal advantage, I will just settle for the former.
 
Last edited:

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi @DataScience ,

Thanks for the reply! Will defiinitely take into your recommendations. And In case you might have missed the qn regarding trying to enter industry directly after AI200 and Kaggle projs vs entering a structured program first (TFIP, AISGP). Which route do you think would be more suitable?
I think after step 2, instead of trying to enter the industry, it would be better for you to rush the OMSCS program first. This ensures that you will be able to get a DS role straight instead of having to work your way up from being a DA first.

You can definitely get a job first after step 2 instead of going through structured program, but I think in terms of efficiency I think just focus on finishing your OMSCS first. If you do well in the program, you will have no lack of job offers. By the time, after you have trouble looking for jobs, you can ping me.
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
It will be likely with the Industrial Phd Program under NUS school of Computing. The other option will be under a hybrid healthcare/data science program such as Duke NUS Quantitative Biology and Medicine. I am currently working in the healthcare sector. I am just weighing the pros and cons vs OMSCS which is a shorter 3 year program hence wanted to make an informed decision before taking any action. If completing the OMSCS is enough for a boost in career progression and if the Phd program is likely to give marginal advantage, I will just settle for the former.
Daniel,

Honestly that program (industrial) is a good one. In fact, I have a personal bias for people with IE/OR background. BUT, if you aim to work as a MLE/DS in the private sector,I feel that it is an overkill, and the additional pay is not going to be worth the extra number of years you spend on the program. OMSCS is more than enough.

You should only seek a PHD if:
- You are interested in academia and wish to seek tenure in an university after you complete the program (very competitive and difficult to get tenure if you are not from a famous overseas university). And the problem is, local universities will not be able to pay you more than what you would get if you were a senior DS in the top tier fims
- Seek ML research role in top tier firm like google,facebook meta in US. There are very few companies in singapore that would be able to pay you the amount you deserve. You'd have to go to top tier firms in US for them to be able to pay you the phd level salary.

In sum, not advisable for phd unless you want to be in academia, or you intend to work as a DS/MLE in US. After OMSCS, you should be able to get a L3/L4 data scientist role which pays more than 120-150k/year. After you work for another 2-3 years, and if you are good, you should make L4/L5 150-250k.
 

idiotical

Member
Joined
Aug 30, 2007
Messages
308
Reaction score
0
It seems difficult to find for real projects, most coursework project like more theoretical and not worth doing it. Any recommendations for projects that ppl can practise on?
 

DDMatrix

Junior Member
Joined
May 16, 2017
Messages
77
Reaction score
41
Daniel,

Honestly that program (industrial) is a good one. In fact, I have a personal bias for people with IE/OR background. BUT, if you aim to work as a MLE/DS in the private sector,I feel that it is an overkill, and the additional pay is not going to be worth the extra number of years you spend on the program. OMSCS is more than enough.

You should only seek a PHD if:
- You are interested in academia and wish to seek tenure in an university after you complete the program (very competitive and difficult to get tenure if you are not from a famous overseas university). And the problem is, local universities will not be able to pay you more than what you would get if you were a senior DS in the top tier fims
- Seek ML research role in top tier firm like google,facebook meta in US. There are very few companies in singapore that would be able to pay you the amount you deserve. You'd have to go to top tier firms in US for them to be able to pay you the phd level salary.

In sum, not advisable for phd unless you want to be in academia, or you intend to work as a DS/MLE in US. After OMSCS, you should be able to get a L3/L4 data scientist role which pays more than 120-150k/year. After you work for another 2-3 years, and if you are good, you should make L4/L5 150-250k.

Hi DataScience,

Thank you for your insights. It gave me a clearer direction on where I should pursue. (y)
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
It seems difficult to find for real projects, most coursework project like more theoretical and not worth doing it. Any recommendations for projects that ppl can practise on?
Hi, instead of finding random projects, I think it would be more beneficial for you to take a dataset from your company, and setup a database with it. You should then find ways to optimise the design of the database schema and create CRUD functions using SQL.

Following which you can go above and beyond by building an ML model with it. Ideate and build something useful for your company The benefit to this is, you can build a POC to impress your boss with, and hopefully cite it as part of work experience in your resume. And perhaps it might give you a legup in your current job as well.

I think this trumps all random SQL projects.
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi DataScience,

Thank you for your insights. It gave me a clearer direction on where I should pursue. (y)
You are welcome. If you are going for OMSCS and have no prior experience in OR, you should definitely take some optimisation classes. It will help deepen your appreciation and understanding of ML later on.
 

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Which bootcamp to recommend to learn more about SWE skills?

SWE = software engineering
GA/Le Wagon.

Sadly I heard the standards of the instructors dropped abit. But they are still the market leaders nonetheless.

Recently, an interesting school "Rocket Academy" caught my attention. I heard they are not bad, you can perhaps check them out as well.
 

Xxeron

Junior Member
Joined
Jun 5, 2014
Messages
5
Reaction score
0
Hi DataScience,

I just finished my NS 1 year ago, now I am intending to pursue IT related courses I have no prior knowledge or experience in this field but I did some research on how to get started on this field. One of the result I came up with was to get ceritified in Comptia A+, Security+, CCNA. I searched it up their courses and cert are quite expensive for me
What is your opinion?

thanks
 
Last edited:

DataScience

Member
Joined
Aug 6, 2020
Messages
381
Reaction score
128
Hi DataScience,

I just finished my NS 1 year ago, now I am intending to pursue IT related courses I have no prior knowledge or experience in this field but I did some research on how to get started on this field. One of the result I came up with was to get ceritified in Comptia A+, Security+, Network+. I searched it up their courses and cert are quite expensive for me
What is your opinion?

thanks
Hi I would need a bit more information from you before I can advice:

- were you from poly/JC and how were your results like
- are you currently in university or intending to go intending university
- The field of IT is quite wide, spanning: cybersecurity, data science, software development to name a few. Which one are you more interested in?
 

Xxeron

Junior Member
Joined
Jun 5, 2014
Messages
5
Reaction score
0
Hi I would need a bit more information from you before I can advice:

- were you from poly/JC and how were your results like
- are you currently in university or intending to go intending university
- The field of IT is quite wide, spanning: cybersecurity, data science, software development to name a few. Which one are you more interested in?
I am from Higher Nitec, result well.. pretty Low I am thinking of actually just doing work & part time study, I have done a couple of Videos and reading so far narrowed down to either: web development, cloud or network.
But as for currently I am self teaching myself Python as I heard is the bread & butter language
 
Last edited:
Important Forum Advisory Note
This forum is moderated by volunteer moderators who will react only to members' feedback on posts. Moderators are not employees or representatives of HWZ. Forum members and moderators are responsible for their own posts.

Please refer to our Community Guidelines and Standards, Terms of Service and Member T&Cs for more information.
Top