3 Types of Data Science Interview Questions

[@jomakaze]

Follow me on instagram for LIVE VIDEOS! instagram.com/jomaoppa/

[@reverend900]

I have 5 years experience as Oracle developer/data modeler. Now I am trying as data analyst(Hive/Pig) etc.. What are your suggestion who are trying to jump in diff. wagons? Love all your vids.. East or West Joma is the Best

[@kp2718]

Joma's insta stories are great! Way cooler than those KZbin videos. Also why I decided to support him on Patreon.

[@兴宣大院君-h4s]

kajima

[@engliyan9158]

Ethan Coulson )

[@firasbouhamdan8603]

Do I have to be smart to be a data scientist?

[@jagicyooo2007]

"Tell me you want me" was the primary reason for rejection.

[@danielprice538]

This video helped me pass my data scientist interview! Thanks!

[@jomakaze]

LOOOL

[@martin.m2]

Fr?

[@silinchen2102]

This video has only been released for an hour. You are tripping.

[@martin.m2]

This video cured my depression though

[@techbro9652]

same

[@sophieblooming8555]

Entertainer trapped in a nerd’s body, or vice versa. The best teacher can make you laugh and think.

[@captnmaico6776]

I once had a question, how the regressor would look like without any coeffecient except the intercept in a linear regression. Could not answer it right away, but got the to the answer after a couple of mins. I asked the job interviewer if anyone ever got it right after immediteatly and he said only one, but he did not hire him. So don't worry question aren't everything, its got more to do with the way you handle the situation with dignity and without being a liar. Also keep your head up, iwrote hundred of applications and still ended up at a big corp.

[@vikasjaiswal2377]

I think the answer is a line parallel to x-axis(considering 2D) since the slop term(w) is zero.

[@captnmaico6776]

technically thats right, as it hold true for any function with only a constant term. But he was asking specifically for the definition of the regressor in an OLS regression. Its the mean of your y's also being a constant, cuase: sum sign ei^2=sum sign (yi-a)^2 -> differiante after a = sum sign yi - n*a =0 sum sign yi =n*a 1/n * sum sign yi=a mean(yi)=a so a is indeed a constant and the mean of your yi's

[@nmmichalak]

I love your videos! I had one quibble about your SQL solution to #1. If some publishers have more videos than others, then their video durations will affect the mean. You could imagine one publisher with 10,000 videos and 9 publishers with 10 videos. In this case, the average duration of that one super-publisher's videos will have way more influence on the mean than the other publishers. I think your question is better answered by first computing an average for each publisher and then averaging those averages. (of course, this detail may not matter in an interview)

[@Melgazar9]

I think this is a good video for suggestions on how to get a data science job at a larger company. However in my experience data science interviews at non-large companies put a huge focus on machine learning/deep learning, linear algebra, statistics, core python programming (sometimes R), SQL, spark/AWS/GCP, and data visualization (tableau, matplotlib, R)

[@xiaoyang8075]

Soooo true. I got refused by many small company in data science but get interviews from large ones cuz I don’t have deep learning skills. The startups wanna takeover the world

[@captnmaico6776]

The reason is simple, its diversity. In a big company workers are far more specialzid and therefore you don't need a data scientist thats able to do everything. Im working as a fraud data scientist/analyst at a big corp and hardly ever do any SQL or spark/aws and stuff related to deploying, because we got business engineers and data specialists, who far better at it and do it 24/7. My task is to evaluate and built the current models and hard blocking scores of fraud behaviour in the system. In a small company budget contraints are far far harsher and tehrefore a good CS guy who is able to perform basic tasks in most stuff and some deep sheet is more efficient for them cost wise. For the people working at startups its usually hell on earth, especially if payment is low and working load big. Been there done that, never again.

[@MoreParksLessParking]

In the above comment, @@captnmaico6776 explains that he is a "fraudulent Data Scientist", who is able to get away with keeping his job because his company is over-employed. Therefore there is always someone else who can do his work for him. He recommends only working for big companies rather than small start-ups, as only they will be able to support this form of parasitic lifestyle.

[@captnmaico6776]

@@MoreParksLessParking sure thing, you know my life and cv do you? In europe people at startups arent paid huge salaries and cant afford basic stuff depending on their location. Its not silicon valley and furthermore, Fuck off dipshit.

[@harshavardhanthorat3910]

@@MoreParksLessParking lmao ded. Do you think he's calling himself a fraud?? Dude works in a project that 'detects' fraud, probably for a fintech company or some bank

[@maximtheconqueror]

This guy does a great job adapting styles from common TV shows to his KZbin channel

[@stanger963]

I got more information about interviews and companies on your channel than all other places combined. You're amazing

[@prakashyadav008]

"when suddenly i felt useless and pathetic." - encompasses many people's life's on weekends

[@wellbrix1808]

Nice video. Thanks. Little mistake @ 3.36 secs with Bayes. Should read denominator 1%x 99% + 1% x 99%

[@KhoiThinh]

I'm sure everyone would love to hear about some materials you've gone through to follow this DS path. Getting an internship is impossible in Japan.

[@jomakaze]

Yea it's a little harder in Japan,

[@neloangelo__13]

maybe try another country I reckon? Unless you are Japanese by birth, getting a job this field is tough as, almost zero chance.

[@KhoiThinh]

In Japan, all the positions they're hiring are almost senior title with at least 5-7 years of experiences so yeah, it's very tough for me. I would definitely consider another country to the pursuit of this career path in the future. Thanks bro.

[@crystyxn]

Wow the STD question was straight out of my A.I. exam!!

[@HeatherHolt]

I’ve been stressing over quitting my current job to finish my studies and finally look for a job in data analytics so thank you for making me laugh while also providing useful information.

[@HeatherHolt]

@code with nikhil well, I quit my job at Apple. Went to school to finish my coding education. Got a job and two months later, COVID came and the whole place shut down. Then I had a baby and now I’m getting ready to start working again. After two years of not working. Life is unpredictable! :)

[@nickn9040]

Makes me so happy that he had to review too lol. 1 year after getting my MS I still would have to review basic probability. E.g. probability distributions, Bayes theorem, etc. Same with the SQL question.

[@crediblesalamander8056]

I only finished my probability course(beginner) a few months ago and I still couldn't answer that easy Bayes's' Theorem question.

[@oleksiifalei2074]

@jomaoppa I think you made a typo at 3:35. It should be 1% x 99% + 99% _x_ 1% instead of 1% x 99% + 99% _+_ 1%

[@utkarshasthana3853]

Thank you! for past 30 mins I was googling a bayes formula that looked like that

[@bigorange6328]

Oleksii Falei I agree

[@kswill4514]

agreed

[@InternetMadnez]

Ah right. I put the same message. 😅😂

[@amawest_]

Ok, good I'm not crazy lol

[@yanrice2]

About to go into a boot camp. Just made this video one of my favorite

[@neloangelo__13]

Lost it at the STD Probability question lmao

[@jomakaze]

Bruh!

[@CGExp]

It's actually a classic Bayes theorem question. Probably in every Bayes theorem book ever written. Hiv is often replaced by cancer in them though

[@ah2522]

@@CGExp Bayes died from an STD because of this

[@alefratat4018]

Let p be the proportion of positive among the population (here p = 1%) Total number of *tested* positive = False Positive + True Positive = 1% of (1-p) + 99% of p Prob of having the disease when tested positive = True positive / Total number of *tested* positive = 99% of p / ( 1% of (1-p) + 99% of p )

[@user-nl9ll1bj9i]

Ale Fratat but if Joma tested positive and the test is right 99% of the time why is the answer 50%?

[@sethnuzum]

Joma and The Techlead collab please, pretty please

[@kubuspat9729]

that would be great

[@TheJFMR]

Joma Techlead

[@chuddywarrior7996]

with sugar on top

[@josecarlosgarcia6477]

This is the best data analysis video I've ever seen. Thank you very much.

[@manojmajumdar3815]

Hi Joma, please make a video on what kind of work someone will get engaged in after getting selected as a Data Scientist, probably tell us what you have been doing lately as a data scientist.

[@MLwithAlva]

LOL very impressive that you can make a video about interview questions so fun to watch 😂 🙌

[@jomakaze]

Thank you thank you :), I try

[@buildwork5177]

Joma Tech bs reply

[@axcelleria]

He looking for girl :3

[@yoosunglee2423]

His videos are my guilty pleasures

[@xypzs5123]

I like how you say you win some you lose some, even though a person of your resume probably got accepted to all of them. Its a very good message. You are born to be a youtuber fr and you take the meaning of the word "influencer" to its supposed meaning. Thanks a lot

[@junaidchoudhary9110]

Man your film making skills is lit AF seriously, keep it up!!

[@anirudhupadhyay695]

This guy is funny!!! and helpful at the same time. Love from India brother!

[@falcxne]

Hey Joma - the new thing now in Data Science interviews is 'modelling' questions (makes sense, really). Curious if you would do a video on how to prepare for that? For many companies, that is the make-or-break portion of the DS interview process.

[@wrcz]

u must be rly cute to succeed at modelling

[@ayoni02]

@@wrcz 😐

[@rana31ify]

Joma you are the casey nietsat of tech community the content you create and the way your tell story awesome, long way to go I have a huge respect to you man wish I could work with you in any way. Keep doing the good work 🙏🏼

[@jerryanyu8467]

Joma, you create good videos for CS majors. We don't need more of "How I learned to code" or "how I got into Google" KZbinrs. Keep it going!

[@andyanderson222]

Man, i just love your sense of humor!

[@ulgk]

I am amazed by your discipline to read/relearn many chapters for the data science questions. Recently, I have graduated from a university with a bachelors in decision science (with emphasis in data analytics). This degree is from the business college. To get prepared for the interview is quite ridiculous in my opinion because you have to know or retain a lot of knowledge. SQL, Python, and R are the programming languages I have to vigorously study. I need to be more disciplined like you. Anyway, thanks for the video.

[@pinsandpans]

So glad that I found your channel so early on. Please don't stop producing content!

[@bedantasaha5699]

I am fan of your channel for a long time now and I am a google product analyst interview in 4 days and I am banging my head instead of something else. This video popped up to my suggestion and watched again to relax a bit. By the way there is a typo in the denominator of the Bayes’ Theorem. Keep up good work!

[@stevengarcia7277]

I am getting ready for my last technical interview , thanks for this video.

[@dineesuka]

a recruiter just approached me on LinkedIn for a data science position that's super well paid and I barely remember how to write hello world in python. I'm just gonna binge watch your channel, maybe I'll figure out a way on how to sound like a data scientist overnight

[@aarzooverma4880]

This is the super cool video I've ever watched regarding the interview questions. Loved it ;)

[@txxxchao]

Great sense of humor!

[@Austin-wh4yi]

Thanks for taking the time to show some actual sample examples.

[@lizhi3003]

Sorry to correct you, you have a typo on STD probability example. Denominator is actually 1%x99% + 99%x1%. In case someone is confused about the answer. But very informative video!

[@foxskinn]

LOL so tru everytime I call my mom she never picks up 😭😂

[@andresr5847]

Hey Joma! I work as a valet attendant at Facebook building 20 and am also a SE student at SJSU. I think i saw you today in the cafeteria during the Japan and Belgium game today. (in a grey hoodie) just wanted to say that i am a big fan and thank you for making funny and informative videos because I’m the future I would love to work up there rather than down here in the parking lot -Andres

[@andresr5847]

*in

[@emilyyang3478]

This is great sum up I had an onsite this week very similar to this structure. And the watching dududu part is hilarious because I was also watching bp MV for hours lol

[@PacMan755]

This is the Joma that we want!

[@battilanast]

Quite like the way you subtly baked an explanation for the a/b testing since there were some requests for that in the comment section from the data science pyramid video

[@tumo6431]

4 years working for advertising agency as a strategy planner , and don't wanna stick on this job anymore. I really wanna to learn more about marketing research. But when I watched your video, you inspired me to discover more data science. I will study data science at Silicon Valley next year. Hmm, I'm really nervous a little bit, but do you think could I become a data scientist when I'm 26?

[@ajieprayoga2434]

Smart channel, subscribed! Keep it up, Jomaa!

[@huangmaurice1672]

I am a total fan of this type of videos. Joma, please make more of these educational clips.💪🏻

[@datanerd112]

Your content is amazing. I really like all your videos about DS. I wish you do more related ML/DS. Thank you so much for making this video, I've been preparing for interviews.

[@partyushkumar4014]

Buddy this is sooooooooooooooo Helpful ✌️

[@justinzint4666]

Thank you for this! I've been interviewing quite a bit lately and this is super helpful. Brilliant was also the first app listed in my app store after watching this. Weiiiiird

[@jomakaze]

ooo interesting... big brother is watching

[@ruslanuchan8880]

Funny and informative as usual! Thanks for the vid Joma

[@jessicasoares9613]

Hahaha, I loved this video! Joma, you are so funny and give such valuable tips! Thank you for sharing :)

[@meihui4761]

Maybe three years late, but typo on bayesian solution. Bayesian formula for P(STD|+ve)= P(+ve|STD) * P(STD)/P(+ve) . We are given P(+ve|STD)=99% and P(STD)=1% (1% people of all of this has STD). We need to find the probability of P(+ve). We know from 1% of people that has std, 99% will test positive. We also know for those who doesn't have STD , 1% of them will test positive. So P(+ve)= 1%x99% +99%*1%. the + typo is fix! plug and chug to give you 50%.

[@2323TheViper]

I really love your channel, it has helped me understand what DS really is. I am an incoming freshmen and will be pursuing DS!! This seems hard right now but as the years go this will all make sense. Keep up the good work.

[@furryannilator4528]

How u doimg rn

[@patsukhum8145]

Really loving your new videos. Keep them coming!

[@jomakaze]

thanks Pat!

[@InternetMadnez]

Hi Joma. You motivated me to get back to maths, data and stuff (I'm just devops currently) Well I feel like at 3:35 you made a mistake: P(positive_test) shoudn't be 1%x99%+99%+1% but 1%x99%+99%x1% right?

[@notfrommoscow4071]

Guys he called Jomama

[@yeshesdevi]

1) Why not just instead SELECT AVG(video_duration) FROM publisher_info? 2) Ok, Joma is tested a second time, and again it is positive. What now is the probability of his having it? But thank you... I think I'd do ok on all three, so I have a better chance than at least the proverbial snowball in you know where...

[@AbdulWahab-oo4fo]

Damn, dude you're quite the hardworker.

[@JGunlimited]

Intro was gold

[@hannahxu333]

It seems to me the calculation using BAYES’s theorem is incorrect: it should be 1% * 99% / (1% * 99% + 99% * 1%)

[@arifhuseynov9937]

Yeap, I was searching for this comment :)

[@bluelinkgames8757]

Holy shit i was searching for this comment too, i just hope were all not to dumb to see something there.

[@Splish_Splash]

yep, i also don't know why Jomma put there + sign

[@mamneo2]

Incroyable.

[@desperado0105]

I think STD test part you mistyped the equation on denominator which should be p(+) = p(+|std)*p(std) + p(+|non-std)*p(non-std) = 0.99 * 0.01 + 0.99 * 0.01 = 0.198

[@kazz5742]

yeah but the answer is still 0.5, you dont even need to calculate that, just take it out as common factor and you can see its a half

[@RahulYadav-nk6wp]

Or simply taking the intersection would do.

[@mg6192]

You are looking for p(+ | std) and not p(+)

[@shreedharraj6777]

Thank you for this video! its very helpful

[@michaelnguyen418]

Hi Joma, is it possible that you start a video on how to start a career as data scientists/analytics? I'm really intrigued with that. Thank you!

[@inchinaxp8663]

Hey Joma, thank you for this. As someone wanting to get into DS, this helped me.

[@randyhoang7264]

Joma, make more of these kind of videos (entertaining interview prep videos). Thankz

[@adoniskomplex91]

"Du" means "you" in german.

[@DarkyMoto]

Same lol i was watching blackpink before i had my recent Data Science interview 🤣

[@Marwin3st]

On the first SQL question, why not just write SELECT AVG(video_duration) ? Btw, Great video, love your content, huge inspiration and I really appreciate videos like this!

[@garrycotton7094]

​@Mathis Coffin Count is actually correct rather than count distinct. For a publisher ID with more than one record, you still need the average video time for that publisher ID. Counting distinct means you're essentially summing up video time for that publisher ID. Easiest example is if there was only one publisher ID but 20 videos. Using count distinct, your average would be the sum of video time. Additionally, @Marwin3st - I think AVG would be just fine, it's simply that some implementations of SQL lack the function.

[@KrushnaBorkar]

Thanks Joma ..That's what I was looking for...😀

[@jomakaze]

perfecto

[@henryalferink1941]

Hey, thanks for the vid. I'm pretty sure the answer at 3:36 is wrong. Answer should be P(test+|has disease)*P(has disease) / (P(test+|has disease)P(has disease) + P(test+|no disease)P(no disease)) = 0.99*0.01 / (0.99*0.01 + 0.99*0.99) = 1%.

[@lpis150]

hi, how did u find out that P(test) = (P(test+|has disease)P(has disease) + P(test+|no disease)P(no disease)) what equation do u use_? thanks

[@tommoritz79]

Thank you for all that great content!

[@avanishpatel4936]

Awesome video content bro 👍👍👍

[@Awesomousprimo]

The intro HAHAH! You're such a unique KZbinr!

[@EmileAI]

Congrats for the 100k !!!

[@wellwhatdoyakno6251]

also I just had to pause, subscribed not knowing who the heck you are ! and never before saw your videos right from the first few seconds of watching your video, I had to pause and subscribe, now I am going back to watching the video.

[@edabbas]

Love your videos 😂 Funny and helpful, thank you brother

[@ayoubed3496]

tick tock tick tock tick tock , i imagine some interviewer doing this to me XD.

[@yeshesdevi]

Or they hum the Jeopardy theme music while they wait for your answer.

[@kamaleshpramanik7645]

This is an awesome video ...

[@khacquyinh3476]

It is really impressive!

[@shashankraman2512]

Watching this video for the fourth time now and still it cracks me up! ;)

[@RamkrishanYT]

We miss you dude

[@CommandoMaster]

Thanks for the tips boi>

[@ashokvj8342]

Ur great jomaa

[@alwardafahd2887]

Hi Jonathan, great video. which software you used to build your resume. looks neat

[@MrPrashantbagga]

Love it! you are amazing

[@ledarkpoet]

I’m so thankful for this video:

[@genealso6726]

Its crazy how different the interviews are for tech vs finance. My tech interviews were so easy but my god investment banks and hedge funds are insanely hard. Wall street really only hires the best but I guess that is why the salaries are so high.

[@williamhans1302]

Hi, Joma. Great video! I have a question, in your opinion, is it possible for someone to start a career as a data scientist while their educational background is not technology or math-related major?

[@thomasgarman6353]

Dude I’m gonna go for it, I don’t care if my major isn’t Data driven, gonna learn over the summers

[@Doon22]

Thanks for another great video. Was worried about the equation, wrongly typed. Though the answer is correct.

[@alexm6697]

Great video... Keep it up bro

[@jomakaze]

thanks alexandros

[@ryansam]

Hey Joma, I plan to move to San Francisco soon and since you kind of recently got a new apartment I was wondering what the options are with a software engineering salary. Could you possibly make a buying guide or even just make a video on how you got your apartment? It's looks pretty nice from what I saw on one of your older videos. Nice apartments that I've seen online seem to be around 3k - 6k mark for like a two bedroom. I think it would be a cool video idea and really helpful for people who are about to start their career.

[@etiwardanis5688]

Love to watch your videos.

[@knnhzzh7924]

for a non-english speaker like me, it is so hard to refresh my stats memory by looking at questions written by English. it is a multi-tasking process.

[@ThinkAloud]

I am bit scared to even choose this field after watching your video.....but my heart says I should at least give it a try

[@synckid]

Bro awesome vid shows allot of effort

[@kctv1193]

hello joma great video!!! please which programming language would you recommend for one to learn.?

[@byronwilliams7977]

I started watching your videos for the content, but stayed cause you're legit funny man. Keep up the great work.

[@turbogt4410]

great video again

[@gouravkumar9011]

Woah this is useful