I Was An MIT/UCL Trained Data Scientist Now Quit DS, Self-Taught As Full-Stack Founder HowDidItWork?

[@SeanTechStories]

To stay in touch: LinkedIn - linkedin.com/in/shen-sean-chen Twitter - twitter.com/ShenSeanChen Instagram - instagram.com/sean_tech_stories Discord - discord.gg/TKKPzZheua

[@matthuang21]

Very realistic take. The data is definitely an important part of decision making, but at the end of the day there are other factors at play including politics at very large companies. You’re a small fish in a huge pond when you work at large companies, so the actual impact you have is limited vs being a founder or working at a startup.

[@SeanTechStories]

Yes sirrr! How are you Matt:)))? On the other hand I was just thinking about this this morning, as a startup, you have probably 10-20 decisions to make in a short period of time, but for a large company there are a couple of critical decisions to be made in a quarter or so. Therefore the decision making process and the weight each decision carries vary too.

[@seththedatanerd1743]

Wow, this video has come at great timing. I am also doing data science at UCL and I have come to exactly the same realisations whilst running my data startup alongside university. I am in the science library quite a bit so hopefully we run into each other at some point!

[@SeanTechStories]

Amazing! See you around!

[@SeanTechStories]

Who’s your fav prof at UCL stats dept?:)

[@GabeIAM93]

I also share your form of thinking. I went down the rabbit hole of trying to understand the underlying theory of every model. Now I shifted more towards a business mindset. I don't have the desire to become a A+ programmer. And especially towards DS the tools are already their and "smarter" people then me are optimizing it. So in regards to a business mindset I took the approach of going through different business problems and figuring out out to productionize a solution that can help solve that problem. Im glad I started practicing this because I found that that is the hardest part. Learning how to plug in ML models depending on the problem is relatively easy because your plugging in and playing with some already built model you just have to understand why and what is the pros and cons for that particular model. I still really don't understand what a "Data Scientist" I just bucket them as a researcher. The industry title as DS i bucket them as a "Predictive Data Analyst really" or "A software engineer that understands Machine Learning to know how to productionize it I thought I wanted to be a Data Scientist but once I realize I just thought it was cool on how you can apply these theories in life. So I took the path of getting my MBA and learning how and why to use Data Science tools and I found taking this approach worked better for me because I started building products right away a learned how DS applies to the simple product im making (GenAI PDF summarizer) I know there is a tool for that already but I just wanted to understand the process of making something like that. Before when I took the DS route I would have studied everything about LLMs (which I still plan on gaining more understanding) and then start building models from excel sheets. However this type I got a basic understanding of what an LLM is and built something in a few days. Thats my approach Based on all I wrote I opted to get my MBA and learn GenAI/DS on the side because I dont necessarily want to build or create the cutting edge technologies. However I do want to figure different ways to apply these existing and new models that will be created in the future. Sorry for the long comment but this video really resonated with me. I would love to chat sometime

[@GabeIAM93]

Sorry I did not proof read this I just typed and sent 😅

[@SeanTechStories]

This is exactly how I felt. Don’t worry about sending long messages I think it’s very authentic! I also couldn’t figure out what being a data scientist really meant back then. Eventually I realized it really depended on what the rest of the team thought what being a DS was - they would ask you for data, for analysis, for models and for reports. Sometimes there was more asking on one thing and sometimes on another. I believe the best way to learn now is what you said - I’m curious about building a tool for summarising PDF and I don’t care if there is a tool out there already. I just enjoy learning how to do this end to end and it is fun. Keep doing that! I think this is extremely helpful for us to see a bigger picture and not feel stuck at the one little task we are assigned with at the moment.

[@GabeIAM93]

@@SeanTechStories Yeah now in the GenAI area its really just plug and play which still in its self pretty difficult but the all the API frameworks out their makes it a bit easier. Also I have an additional question. Im starting my MBA in Jan and curious on your thoughts on a focus. Im thinking Finance & Strategy, because I do like learning the End-to-End on why companies chose different projects/products. All the "AI" stuff im learning on my own. What are your thoughts ?

[@SeanTechStories]

I think if I were to do an MBA, I’d try to connect with more people in both business and tech world. Doesn’t matter if I’m majoring in business, finance or AI/tech, as long as I keep myself up to date, I’ll be fine.

[@Simba365]

This in all industries where you dont do any cutting edge stuff and dont use the majority of what you learned in school. Honestly outside of research every industry is like this

[@darshantawte7435]

I totally agree with this video. I think because of these shitty bootcamps and even some college degrees have glamourized DS/AI/ML, and everyone thinks they can get into this field. I feel DS/AI jobs are very exhaustive (currently working as an ML/Cloud Engineer at a top MNC), there is always that sense of incompleteness (which you won't face in software engineering or other tech domains), hence people who are naturally curious and have that experimental mindset combined with deriving actual value to a business thrive. Also the requirements vary from company to company unlike a typical SDE job, and most companies nowadays expect you to know all the SDE skills on the top of AI/ML for a job that pays even less or at par with an SDE. With the advent of LLMs and powerful pretrained models, the model training part is already taken care of , productionizing , scaling and deployement of these models is the next thing to focus on as they models are used by millions of users, so Devops/MLOps is what i feel is something which will be in immense demand going forward.

[@SeanTechStories]

I think what you described is very accurate - the capability of deploying fast and iterating is the key. Thanks for sharing your thoughts!

[@akmalsaputra2020]

so devops, mlops, cloud, and security is the next big thing?

[@SeanTechStories]

Not too sure, but I know a lot of MLOps startups have pivoted to LLMOps now. I guess for any career it’s important to keep iterating

[@MuAlex]

Feeling super connected when you mentioned switching from ds to pm. Exactly me now 😢 don’t know how to put it, it just feels interesting like writing those user stories and figma design but also weird and uncertainty to some extend.

[@SeanTechStories]

Yes, I feel you!

[@danielfayoyin2409]

Bro, as someone with over 7 years in this game and contemplating the next steps. This has been very helpful. Nice to have some validation of my experience. I really felt hoodwinked after my first 6 months in my first job. Good luck with what you build and create! Very much pivoting on my end also

[@SeanTechStories]

Hey Daniel best of luck on your next steps! Thanks for sharing your feelings!

[@RedSky8]

I've been trying to enter the data science or software engineering scene for a bit. It's hard with all of the great people also trying to get a job, but thanks for your take on the differences in the roles. I think I'd like to stay on the cutting edge of tech and do more exploratory data science and development.

[@SeanTechStories]

Staying on the cutting edge is the best way to be flexible and agile😁 Good luck!

[@junda2134]

I entered data science industry with no prior experience a half year ago when I was 27. This video was so inspiring, thank you!

[@SeanTechStories]

Thank you!

[@max_7344]

Do you still believe it’s worth sticking with data science?

[@SeanTechStories]

It’s worth being multi-skilled and not sticking with any one skill set.

[@shenseanchen]

@@max_7344agreed with your example! Best of luck on your next steps Max!

[@max_7344]

@ How do you think the outlook for the cloud field compares?

[@davishandler4503]

Great vlog. Keep up the good work.

[@SeanTechStories]

Thanks Davis!

[@Artonox]

Share similar values with you. Im much older than you, not a DS or tech person, but was at Maths in UCL, and I thought could use it to help bring new innovations. However all the finance tools are already built, ended up as an accountant in an unfulfilling career path. Money is good though but im stuck to change in a golden handcuffs way.

[@SeanTechStories]

Hello UCL alum! I used to Gordon St a lot to submit my math homework haha! I totally understand when it feels like we are stuck to change with golden handcuffs. I used to have a pretty decently paid job in tech and I felt stuck coz I felt like switching to founder mode or software would either take a lot of resources/money or a degree. But AI changed my direction completely. I felt like it empowered me to imagine any potential I could reach and could guide me the way to it. I feel like it’s always hard to change but at least now we’ve got some new methods to change.

[@CatchingOnFire]

Sean! thank you so much for making this video. I see many reddit posts and pieces around the internet indicating that data scientist end up disappointed (feeling like they clean, maintain and write sql most of the time)... But because so much of KZbin has videos talking about data science hype from people with the intention of selling a corse or book affiliate link, its been hard to get a clear picture of if the field it truly right for me. But this video and all of the points you made are SUPER helpful because you take it from the perspective of the big picture of the role. My only question: You talked about your experience learning statistics and pointed out that a lot of the statistics and problem solving you learned not utilized because its already solved(?). Im currently about to start my first semester of college and embark on the same journey you underwent... Do you regret learning heavy statistics and data science skills or do you feel like its still good to learn these?

[@SeanTechStories]

Hello my friend, thanks for your comment! I'm really glad this was helpful for you. I agree that there are a lot of contents online encouraging people to take DS or Data Eng roles because they can sell their courses. I personally feel quite grateful to a lot of the contents I watched about DS on KZbin (I never paid for anything though lol, only free content on KZbin and read a lot of blogs on Medium) because those contents helped me prepare for my career. But what I ended up realizing was that those preparation was mainly for the interviews and how to get into a job, but the reality of the jobs often times can be a little disappointing, from a big picture perspective. But though all of these being said, I do have friends who enjoy DS roles so please talk to more people and make sure you make the best decision for yourself! Regarding your question, I actually do not regret what I learnt at school. I think those were really at the forefront of ML research at that time (Deep learning, transformers, all those classic machine learning models etc). This being said, I think you should think about what is at the forefront of whatever you choose to learn right now, because things can change very rapidly in the next 1-5 years, so the only way to make sure you are not behind the trend is to keep yourself updated. Back then we were still learning SVMs, LSTMs, which I think are quite outdated now because people just basically use XGBoost or LLMs for certain tasks. Also, think deeper about what careers you might wanna take, and talk to those people who currently do those jobs you dream of (find them on LinkedIn, X.com, etc), and ask them if they actually end up doing what they dreamed of doing. I think this is the most up to date information compared to what people on campus, youtube, or your professors would tell you.

[@CatchingOnFire]

@@SeanTechStories Thank you for the advice. I love the idea of reaching out to other people in the field, and will definitely seek to learn more from other people in the Data-Science to understand whats right for me. I kinda have a problem... I notice that I don't like the idea of cleaning and structuring things for a job so Data Engineering/Science may not be for me... But at the same time I also don't see myself as a builder- because although I LOVE to innovate, usually in order to build something it requires a lot of planning and organization which is kinda a weakness of mine 😅 But I probably just need to work on my weaknesses. I'll figure it out eventually lol Anyways, I was mostly curious on whether or not you feel like your background in Data Science and Statistics skills are helpful even though you are not a data-scientist anymore. I haven't gotten to watch your other videos yet (so I might find the answer there) but do you feel like this background gives you an edge over other entrepreneurs?

[@SeanTechStories]

@CatchingOnFire to your question, yes. It’s helpful in a sense that you need to think ahead of time before your product grows big: what if we have a lot of users, what do we need to track when we set up our data base (user clicks, sign in time, usage time etc etc). And if we need to provide some smart features in an app, like if we need to recommend something or predict something, what data do I need to collect in day one when a product is still just an MVP. All of these come from the sense we developed as a data scientist.

[@Californiansurfer]

Product managment : you can use. 1. Human judgment 2. Human intution You use the data to make decisons. Again, the human is always in charge … We are responsible. ❤❤❤

[@SeanTechStories]

Yes

[@edwardlin2941]

Thank you

[@SeanTechStories]

You are welcome.

[@brentdixon1831]

Sounds about right. Although I do get the sense that 2 years is not so long and maybe there's some impatience here and a lot of expectations about making an impact. I imagine this guy is probably at the top of his class and very smart. But I think personally, to make an impact you need to talk to the people around you and advocate for the change you want to make or incorporate the feedback of a lot of others as well, I don't hear a lot about this being talked about here. Mostly what I hear is, you can't make an impact and most corporations are not so data driven as they tend to say. But I dunno, you know, how much impact can you really make other than funneling more funds into shareholders pockets? That's always the "bottom line" as they say, having a monopoly and changing the law and sabotaging the competition. If any company would honestly just care about their product or customer more than anything else, or the positive impact they could have on the world. Edit: you should look up that Japanese dude who invented the blue led and see what this guy had to go through to get it done.

[@SeanTechStories]

Hey man! Fair point. Btw this guy was not the top of class haha and always felt imposter syndrome at university while there were too many extremely smart kids around me. I agree with you regarding having more patience and talking to people around us and advocating for impact. However, I would say those who ended up advocating themselves well, though indeed were more patient and did great jobs at promoting their work, spent a lot of effort focusing on how to manage corporate relationships well, how to ‘not disrupt’ the existing ways of working so that they can get their ideas through and execute. It’s a personal choice imo, some of my previous colleagues really enjoyed spending years in one role and ended up becoming the influential figures in the team, but some others like myself, prefer to build fast and iterate fast, most of whom eventually turned to the startup world. I’m not trying to convince everyone to follow the same path as us, but I guess my opinions resonated with quite some people who watched this video or who might feel the same struggle of not being able to make impact or feel stuck. I’m providing a perspective for people who felt stuck. In fact, I talked to quite a few of my colleagues in DS who spent 5-10 years in such roles, and the complaints were very similar. Most of them either accepted this as a reality and start to play with the game or they didn’t have energy or momentum to push for changes any more.

[@arham2916]

Hi! could you elaborate on the type of roles graduates are moving to in finance. I'm currently doing my Undergrad in CS and always wanted to find some road in between which be a good mix of CS and Finance. I was also contemplating doing my MS in DS but you've really made me think about it a lot more critically now so i'd like to say thanks. Good luck with your endeavors, i understand that it was probably a tough decision.

[@SeanTechStories]

Thanks for your comment! For finance roles, what I’ve seen are quant developers, quant analysts, data scientists, etc. My friend who works as a QA/DS at a hedge fund told me he had to do a lot of Jupyter notebook data cleaning and eventually just plotting a lot of charts for presenting a conclusion, which was fine as a job but sometimes he jokes about why he’s spending hours just using these basic pandas or pyspark functions. I feel like perhaps a good mix of CS and finance jobs which are fun could be engineering roles at some fintech startups/corps. You get to see the products you are building going live and maybe need to do quite some firefighting during tough times but at least you feel ownerships (i might be imagining here coz I haven’t done it before). And no worries! Talk to more people, especially practitioners in the industry, reach out to them on LinkedIn or X, and just ask them what their jobs are like, you can get a much clearer picture on what the reality is VS what KZbin and blogs would tell you.

[@mezzela00]

hey mate thank you for sharing your insights. I also was frustrated to see that coming from cs, i done the exact same basic things as my collegues does wich they came from non technical areas like econ etc. but it is what it is i guess. can you tell more about your product managment journey? daily task etc. thanks in advance

[@SeanTechStories]

Thanks for sharing! I can def share more in the future videos and in fact I’ve made some videos about my product management experience in the past as well. Generally speaking PM role is much broader than DS. We get to work with UI/UX, Eng, DS, Biz and all. But you kinda become a doc writing machine and you are constantly on zoom calls. I feel like it suits really well for people who like managing a team with influence not authority, but for people who still enjoy building things handson, it could be a little too much talking and less building.

[@Hab2320]

Hey, I’m doing sql hell. I studied physics and economics in uni… It’s annoying af. Where should I look for next for jobs in the world? Love math and tough problems, strategy etc. any clues? Also how do you feel about startups…? Any desire to do that?

[@SeanTechStories]

I think if you love tough problems and strategy, startups are the way to go:) Writing SQL is a waste of time.

[@stevenayare6097]

What do you think about using Reinforcement learning and Rnd Jobs in Finance?

[@SeanTechStories]

I don’t know. But I think for any tech it’s important to see how long people have tried to productionize it and create business value with it. If RL has been widely adopted in trading nowadays then I think perhaps the foundational work is done, but if it’s still remaining to be in R&D, it could be cut off in the future or will take off wildly. You need to take a bet.

[@2X-.-isSerious]

Is it realistic for someone to be good at both ML/DS as well as SDE? I have immense interest in both areas, but from what I've been told, specialization is key to get a job in this market. I don't want to fully commit to one, I love data and I also want to be someone that can build an idea into a product. What would be your advice Sean? I'm in my last year of CS undergrad, i feel a bit lost.

[@SeanTechStories]

Hello! Yes it is realistic. Before LLM, I’d say try to specialize, post-LLM, I’d say learn as many things as you can:)) I think the good news is hat now you don’t have to choose either, you can be both because AI basically reduced a whole chunk of memory required for our brain to remember all the syntax for the getting things to work. Setting up environment and getting things to work in production have become incredibly cheaper now. So you might see some startup teams with one person equal to a whole full stack team + data scientist. I’d say don’t set any limit to yourself at this point. For jobs, it depends on what attracts you more and what jobs are available. These days massive layoffs are happening for both roles but also loads of people are building new startups. I think it’s cool to join any role to start with but please ensure one thing - you are constantly learning and trying new things.

[@2X-.-isSerious]

@@SeanTechStories Thank you for the response Sean!

[@SeanTechStories]

@@2X-.-isSeriousno problem!

[@prateeksarin3555]

Very good

[@SeanTechStories]

Thanks!

[@SeanTechStories]

Thanks!

[@maheshbanshiwal3274]

Hi sir...I don't know what to do in my life I trying to learn data analyst but I do not know should I really learn it or not .. because some says it is a end career because of AI...so I don't know what to learn and get a good salary job please sir guide me P.S. PLEASE DON'T MIND MY ENGLISH I KNOW IT'S NOT GOOD

[@SeanTechStories]

Hey Mahesh, don’t worry you explained yourself really well in English! As a second language speaker myself, I appreciate anyone who tries to communicate with another language! I think working as a data analyst could be tricky in the next 5-10 years. As you mentioned AI is gonna significantly accelerate the speed of writing queries and building models, so perhaps one really good data scientist can do 10x amount of work compared with pre-ChatGPT era. But that being said, I think having a good sense of business is the key. Tools are gonna get cheaper and easier to use, but that also means we are being relieved from the need of grinding on doing the work that is repetitive and unproductive. I think for getting a good salary, it’s always about how much business value you can bring in the job and how rare the skill set is on the market. I don’t have a good answer for this to be honest because I don’t even know how AI will evolve in the next few months. But I do know it’s a great time to do something EndToEnd because we are enabled to build and sell end to end now due to the new tools. If we don’t do that, our competitors would. So I guess that’s my genuine thoughts, which could be biased.

[@Srini-v6j]

What do you think of data scientists who are working in research? say in public health domain?

[@SeanTechStories]

There could be more exciting work there. I’m not entirely sure though.

[@HelanDeC]

谢谢你的分享！我梦想成为data scientist, 现在是BI consultants. 想通过data engineer job 和data scientist 一起工作😄

[@SeanTechStories]

加油！

[@dli4268]

why you moved from states to UK?

[@SeanTechStories]

To get some freedom to explore without visa restrictions:))

[@stan188]

that's right human judgement is more ignorant - invest in TSLA 15 years ago, who needs to work now?

[@lam-thai-nguyen]

I had to watch this video twice to understand what you meant. I’m not sure if I’m your target viewer but I too don’t want the kind of “writing SQL 80% of the time and the decision is made mainly based on human judgement”. The way you’re telling your stories sounds like a leader and a great adviser to me. I want to ask you some questions. From WHEN did you realize you wanted to do something bigger, not just what a data scientist does? If you had started what you are doing now (startup) back then, where would you be now (without valuable insights into professional working reality and realization along the way)? And at which moment you knew you were gonna quit? Good luck with your startup.

[@brhnkh]

Yeah he is a great communicator, definitely leadership material.

[@SeanTechStories]

Thanks so much for your compliment! To your question, I guess I always wanted to do something bigger and I love exploring new things. Whenever I bump into obstacles for further exploration, I will realize this is not for me. To give you an example, I normally would love to reach out to VPs and Directors of any org I join, and book a call with them and try to understand from their perspective what problems they believe the team are facing. I thought this was something a data scientist should do because we need to understand what’s important and we need to be unbiased by referring to diverse information sources. And most managers I had were supportive but one time one of my managers at a former company told me I should not do that because that is the manager’s job to speak with Directors and VPs to align on goals and strategies. My manager told me my job was to execute what was given to us, at that moment I realized being a junior DS was pretty minimal to the team. The senior leadership of course would be willing to chat with us but at the end of the day the expectation for junior DS folks were really low - just get our data right and maybe show me some cool modelling results as a reference. I moved on from then on. If I hadn’t experienced all of these professional working reality and jumped right into startup world, I probably will be more naive to think people working in FAANG are all geniuses but in reality most of the people are doing just a day job. Of course I have met some really talented colleagues but I can ensure you that the work and responsibilities a startup founder would take is significantly higher than a normal role at FAANG, which will make you learn much faster. But, it’s difficult to hypothesise where I would be if I hadn’t had these experiences. And I really appreciated the fact that I experienced large company life and that only gave me more motivation to set no limit in my roles and careers.

[@SeanTechStories]

Thanks a lot!

[@ohyeah4308]

Any career is a scam; be your own boss.

[@emmanuelameyaw9735]

Every job is like that, dude. You are not going to hugely change the world as an employee.

[@SeanTechStories]

Indeed;)

[@SeanTechStories]

Follow-up Video - University Education Is Not Ready For AI Job Change WhatCanYouDo?kzbin.info/www/bejne/g2fQiat8pJmrZpI