• ankitrathi

Your DS Story S01E04: Ksenia Galkina (Freelancer)

‘Your DS Story’ is my attempt to bridge the gap between data science professionals & data science aspirants. Here new crop of data scientists will share their experiences, struggles, achievements & their advice so that data science aspirants/enthusiasts can learn and get inspired.

Ksenia Galkina, as a young and ambitious Data Scientist and AI-driven Enthusiast, just can’t keep calm! She is learning and working on the development of Deep Learning algorithms to solve real world problems . She believes in science and power of the community work for any kind of Industry and strongly convinced: it is possible to learn, achieve and master anything in this world, if we truly put our maximum effort into it. While being open minded, persistent and curious she wishes to innovate and brings whilst learning from others by way of a continuous exchange of ideas.

1. Please tell us a bit about your background?

I came from political science and besides the fact that I used to visit courses from Business and Finance, I did not have more than quantitative methods and something that could look like statistics during my degree — so entering the world of data science was completely out of the blue for me. Especially when you first tried to use ICQ 10 Years ago and that is where my computer usage ended.

I used to work in a lot of different industries. Now I have 5+ years of experience while being 25 years old. I think my first contact or kinda of introduction to data analytics happened by accident, while I was working for the real estate firm in its early stages: when you don’t even have a fully furnished office, when you do any kind of tasks. This can be sometimes destroying, but sometimes very useful. Once I wanted to know the development of prizes in Vienna and I tried to figure out how to do that — this brought me to predictive modelling. And at the same time I started my own side project — a recommender engine for real estate, that I was desperate to implement at work. This all was still kinda “google this”, “try that”, I was not taking an courses — just learning by doing.

When I first started in IT department at Telekom firm there was no chance to touch the code, moreover I was the only person who could program (I already took some courses and was familiar with some machine learning algorithms). My work was very routine and included a lot of contact with data — huge amounts of data to validate between roaming partners worldwide and as we, humans, are all very dependent on our body condition and other factors, the accuracy was not that high. I mean, you can be just tired after working 8 hours straight and starring at 2 monitors with numbers. So I came up with an idea to automate this process, first by dividing the whole process into several problems and try to automate each one of them. While working on this idea, I was so obsessed by bringing it into life that I spent a lot of time learning Spark Streaming, databases, deploying the models, containerization — the whole amount of information and different aspects that data scientist has to be aware of.

This led me to my next position — 1 year contract program inside of A1 where I worked as a Data Scientist and AI developer (whatever that means 😀 ) There were different projects from Augmented Reality App to the Recommender System for Video on Demand that I was working on. Telekom is a pretty much gold mine full of data, that you can use in different ways, but for this you also need a great team of experts from different domains to figure out how and what to do with it.

I have my own side projects, I am very much into deep learning and reinforcement learning and their applications in healthcare. I left Telekom and now providing mostly AI consulting services and working on my website, where I am going to launch a series of articles about statistics, programming, math and machine learning for anyone, so after that you have skills to reach at least Junior level. Also this brought me to the idea of opening data science and AI boot-camp in Vienna, that we don’t have yet. Besides all of these I am co-leading and managing 2 communities : GRAKN.AI and Data science Initiative at the University of Vienna, where we invite tech leaders from different companies — mostly startups — to give talk and workshops to show how the “real world DS works”, we also organize hackathons for anyone who wants to learn programming. So this is plenty of work to do, but also a lot of fun.

I guess, the fact that we all have it here, so much data, we just have to find a way to get information from it. Especially it fascinates me when i think about Biology, Genetics — Medicine in general — we can do so much with this, process tons of information in a short time, was shorter and faster than any human would be able to do.

2. What projects you are working these days?

I am working on Prediction of Earthquakes . Link is here.

Can’t say it is a business problem, but the result is deeply needed.

*Business problem I was solving was the one described — automate data validation and testings between Austria and its roaming partners, in order to provide error free connection. Machine 2 machine and telephony.

There is a lot of seismic signals data from the timing of laboratory earthquakes. The data comes from a well-known experimental set-up used to study earthquake physics. The acoustic_data input signal is used to predict the time remaining before the next laboratory earthquake (time_to_failure).

I have never worked with this data before and it is always a challenge because you have to understand domain to ask right question and to get.

3. How your day to day job looks like?

It depends. I guess, I am not an office person, means not necessarily pajamas and bed, just good atmospheric places with lots of light. Also i am travelling a lot and airports are sometimes awesome to get some shit done.

80 % Data import + Cleaning + Feature Engineering

20 % Modeling, Tuning

This can vary, because at my previous job, I spent around 60 % of my time talking to people from different departments, advising on the data i am working with with their domain knowledge.

4. How you started with DS or transitioned into DS?

There are so many resources out there, too many! Like seriously, sometimes I think it’s not a good thing, but the best side of it is in the fact, that lots of them are free and just great, very interactive. The problem I had — I have never ever in my life coded or touched computer science, so I was overwhelmed and alone on this way, this is what brought me to the idea of visiting meetups and one of the best things was Lemmings incubator for AI and design. It gave me a lot of inspiration and community feeling.

Also, after lots of reading and researching you forget that programming is all about practice and applying this knowledge, not just theory . So I was scared for a while to try things out and just watched and watched videos and tutorials. A really great help came from Kaggle Learning. You can try out everything right there in terminal and they have bunch of courses now, from SQL to deep learning.

I decide what problem I am trying to solve, thinking about the time I want to invest, picking a course. Applying everything that I just read about. Going to the Meetup after that and just ask people about their opinion regarding the problem.

5. What advice would you like to give to DS starters or DS transitioners?

What to do:

  1. Statistics, Math, Data structures and algorithm are your best friends. Udemy is actually very good at explaining complicated things simple.

  2. Less is more. Don’t take everything, every course at the same time.

  3. Try to solve problem form the domain that you are interested in Politics ? Great, take a data set and find a correlation between loving oranges and voting for Trump! Music? why not to analyse people tastes related to the places the places they live? You can even analysis religion if you want, everything is data!

  4. Learn Git and GithHub Art — very useful, always.

  5. Start blogging, I always say, the best way to learn something -try to explain it to someone who has no idea, explain me like a I am 5 year old style. It helps to understand what you don’t know and what you do.

  6. Find a job you want to have. Don’t be shy, Google, Microsoft, Amazon are all the places to everyone. Analyse the positions and structure your learning path this way, you don’t need a PhD, MS or background in computer science for that. Great podcast from Free Code Camp — very inspiring and realistically applicable.

What not to do:

  1. Becoming a Theory Guy, focusing on articles and papers, instead of applying simple small things immediately.

  2. Moving on this way alone. Communities are great, from inspiration to the tips and advises.

  3. Trying to avoid math and thinking about it as a scary hard cracking nut. It can be very easy, if it well explained and you see it this way. Resources and educational videos now are just great, you can get anything after watching 20 min video ( just go and apply it!)

  4. Thinking it is not ”yours” — complete bullshit. You can learn and master everything you want.

Thank you for reading my post. I regularly write about Data & Technology on LinkedIn & Medium. If you would like to read my future posts then simply ‘Connect’ or ‘Follow’. Also feel free to listen to me on SoundCloud.

#DataScience #MachineLearning #TellYourStory

1 view


T: +91 9891XXX969  

Follow me

  • Facebook Clean
  • Twitter Clean
  • White Google+ Icon

©  2020  Ankit Rathi