Dichotomy Of A Data Scientist
Two contrasting behaviours
This is the story about the behavioural difference between an MNC and a startup data scientist. When I was working in an MNC, all I wanted was to work on cool projects — Deep and reinforcement learning. I didn’t care whether the ideas would work or not.
We used to have brainstorming meetings and I proposed ideas which might have taken more than a year to come to fruition. I had hints from my seniors that projects with low risk and high impact will see the light of approval. But I didn’t care. All I cared was about my own learning.
I knew the company is a $4B giant and not going to die if I waste their money. Why should I care about their ROI?
Note: BTW, those ideas had huge potential and there are startups working on them
When I began working in a startup, my behaviour changed completely. My CEO would ask us how can we do this cool thing called x and I would request him to focus on this less cool thing called y because it will be less risky and highly rewarding.
As a recently joined data scientist I constantly thought about — How to make something that works for sure and help us grow the startup!
The reason I joined a startup was to learn fast and also see the hard work come to a good end. I definitely don’t want to put my 100% and hear one day about our closing down.
P(action | startup success)
Hence, it has become intrinsic to me to constantly do P(action | startup success) and not just worry about using cool algorithms.
As I thought and experimented more in my past, I have found that participating in an external competition is one of the best ways to quench my thirst of trying new things. I don’t need to take a toll on my company for my selfish gains.
BTW, I am currently participating in SEMEVAL conference competition and enjoying a lot 😀
My first months
One of the first things I did when I joined the startup was to write a scraper. This is not something you would expect of a senior data scientist. If this task was given to me in my MNC job, I would have easily handed it to someone and relaxed 😬
I think that when a large corporation imposes structure of work, you start exploiting the structure in your favour.
Recently we did a POC and then I refactored code myself, created APIs, a demo with UI(something I never did before) and then deployed it to AWS. I felt the thrill of making progress with each small step.
The whole process was definitely unstructured compared to an MNC but somehow much more fulfilling. When a company starts to segregate work for efficiency, it devoids the engineers from the pleasure of owning things.
Now my focus has changed from working on cool thing to how to increase the value of the product. We have weekly discussions where we talk about user behaviour and the data science for it — search and recommendation.
I never had this opportunity in an MNC because there was a product manager who would take care of converting user needs to features and let us know. I always felt like a backend engineer and lacked a holistic understanding.
It has been 3 months and I am very happy about the change in my mindset by working in a startup. I feel all data scientists should try working in different environments.
The grass is green and different everywhere 😉
Want more?
Subscribe to Modern NLP for latest tricks in NLP!!! 😃
