{ "media_type": "text", "post_content": "How far in data science can you go without a PhD or even solid math background?\nI was never even taught linear algebra but I've been doing DS R&D (basically very fancy feature engineering) based...", "post_id": "5e4cd71c52b490002156f830", "reply_count": 13, "vote_count": 3, "bowl_id": "5ace5097be6e660013abcc42", "bowl_name": "Data & Analytics Consultants", "feed_type": "bowl" }

How far in data science can you go without a PhD or even solid math background? I was never even taught linear algebra but I've been doing DS R&D (basically very fancy feature engineering) based...

like
Posting as :
works at
You are currently posting as works at

i run a 60m analytics practice with a ba :)

like

on and improving on academic papers, for a client problem. I've been able to get away without the deep math stuff cause I'm a good coder and I can quickly understand the intuition (and algorithms too) of a lot of ML so long as I stay away from the actual mechanics of the math. I get lost when we go as deep as stuff like the difference between the different ways PCA is calculated (which, I know, isn't that deep) or how exactly deep learning *really* works. I've been able to do good work but I think I have some serious imposter syndrome when looking at exit opportunities for AI/ML

like

👆🏻 this. You can’t really understand the full functionality behind an algo without the math. I was told in my data science career that you should never use an algo unless you understand the math because they can too easily be misused especially if they are preprogrammed in some capacity. But I will say that a PhD is not the requirement to understand the math.

like

The math in applied ML/DS is not very advanced, mostly college freshman level calculus, stats, and lin alg. You can very easily make up for it with a couple of online courses. Source: was math major in college and did MS in CS/ML.

like

The PhD is only *necessary* for a small fraction of employees when you really need something to work and can't afford a pilot. When you need theoretical justification with guaranteed error bounds etc. Or when you have a true research team which has a need to employ theorists. In practice, businesses don't run this way. A pilot is enough for almost all situations and pure research is extremely risky as monetization goes. Most don't have a need for PhDs who can casually use terms like "Hilbert space." Learning to read the rest of the math and even experiment is most often attained via mathy masters (usually) or bachelors. Don't have to be a physicist or applied math major either, softer ones like econ or analytics will do. Pure CS is a tossup since someone could've "skipped the math" in that, though there are advantages for cleaner code and deployment there.

like

Problem formulation and related statistically valid data preparation is generally a business analytics / product management problem and does not require a PhD. When it does, just any PhD won't do: don't hire a physicist to work on agricultural products when you need that level of domain expertise. You're right that doctors have a tendency to expect special treatment more though, which is a reason against hiring if anything.

like

True. Not required. There are now math degrees focused on understanding algorithms, and many computer/robotics engineering degrees delve into them enough to gain from self study.

like

There's all these schools like Galvanize, Flatiron School, and a few others that claim data science specialty. However, they say you do not need a Ph.D. to apply but the audition interview processes are rigorous enough to require it. I required data science for my program because at the end of my masters I knew I'd need it. Those extra data science schools are appealing if I ever get an honest permanent employer that's not attempting to throw a bunch of ego at me to afford it. They might go beyond current skills. I like how many of them bring in employers and allow pay as you earn while they do their own recruitment. They have different associated employers too. Galvanize is directly related to Elon Musk businesses and it's either Flatiron or Thinkful that is directly related to both Boeing and Northrup Grumman. While I'd like my AI monitored and controlled dam to come to fruition, the fish population counts and some of the diving tourist attractions in Spain mean they are a great start for sharing knowledge with Turkey about how to administer their upcoming dam, so instead of those routes for the ML, I agree with Elon that it's time to start working on some AI policy, including finding avenues for case law. In that regard, I don't need it except for SQL or Python inquiries into the case law, which I still don't need THOSE schools for it since I know that on my own studies :). I have a few more case law agendas to work through. Dam, that guy is now looking like an ego jerk, I am really starting to like Elon, despite long family histories. I'm way better than all that, lol.

I couldn't agree more about the area-specific stuff. I have had multiple experiences with law firms, big and small, and in pro bono affairs. I've had courses in law, administrative justice, and regulations. I also have experience preparing permits, writing permitting manuals, and writing policy. My Ph.D. is in Business Management with a leadership specialization, so it says leadership on the degree. My expertise is needed for data extraction, evaluation, reconciliation, interpretation, and strategizing. So, creating a new department at a law firm, starting a metrics program, and generating strategy to more smoothly grow a law firm. Its also good for politics and designing prism for locating additions to case law to use for developing policy to transfer into politics. However, if the goal is just for reporting dashboards unless you need to interpret for anomalies then investigate the anomalies, it's probably less expensive to hire someone with fewer credentials. However, if it takes multiple people with fewer credentials to do what you need and a Ph.D. can do it themselves and possibly more, it names zero sense to go the extra labor route since that ends up way more expensive.

All that being said, I wanted to add ->. I am not good at plant identification beyond Brazilian nuts and orchids, a squirrel is a squirrel unless it flies, cows can have as many shaped spots as they they, they are just a cow, but horses, you know what??? That's a love story :)

You’re fine for any type of strategy role but depends on what you want as a exit

Additional Posts

Not seeing any other posts!

Testing for sound!

Ahh it is posting but into middle of feed as opposed to top.

like

Some Asian girl just generously offered me her middle seat for my aisle seat. I kindly declined.

like

Resigned and they set up a meeting to get me to stay. I dial in have been waiting for over 20 minutes. Burning bridges now. 🔥

like

Just paid rent for my apartment. Experienced a high like never before. Totally maximized the value!

like

Can't help but wonder how clients are functioning and operating. They can't even do simple excel! It shows that it doesn't take much to own and operate a profitable business I guess

like

currently taking applications for a travel buddy. role description: hang out with me when my company sends me to countries where I know no one. qualifications: preferably a male, attractive.

like

Seems like a bunch of amateurs on this app lately.?s like= How much salary can I expect? How do I get a new job? Should I wear makeup? Has me feeling like a genius. Answers= Glassdoor. Networking.Yes

like

Does anyone know where I can get cheap land to dig a hole and crawl into? Doesn’t have to be domestic as this is the end of my career. Thanks!

likefunny

Hi, who joined?

like

Is it possible to negotiate the salary during performance year end at Deloitte?

like

That feeling when you walk past the long security line, hold your phone up to TSA agent without stopping and breeze through pre-check.

like

Analyst emails client: "sorry, can't come in today, immovable personal conflict". Client responds: "Good luck on your interview!!!" 😂

like

What's your Consulting end game (Entrepreneurial, Partner track, VP industry, etc.)?

like

Favorite and least favorite US cities

likefunny

Deloitte is one of the few major, publicly-listed contributors of Trump's Inauguration. 😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡😡

like

When you have an aching feeling that you could be doing more to get ahead in your career but are burnout at the same time #deepthoughtintheshower

like

When you find a 10-year old flip phone in your childhood room and think back to a simpler time

like

30% bonus 20% salary increase plus promotion. Boom!💵💵💵💵💵

like

New to Fishbowl?

Download the Fishbowl app to
unlock all discussions on Fishbowl.
Download Fishbowl to see what others are saying
That was just a preview…
Sign Up to see all discussions
  • Discover what it’s like to work at companies from real professionals
  • Get candid advice from people in your field in a safe space
  • Chat and network with other professionals in your field
Sign up in seconds to unlock all discussions on Fishbowl.

Already a user?
Login here

Share

Embed this post

Copy and paste embed code on your site

Preview

Download the Fishbowl app

Get realtime notifications and see
what's happening in your industry
from the palm of your hand.

For account settings, visit Fishbowl on Desktop Browser or

General

Legal