Alumni Interview # August 2020
Welcome to the august edition of our alumni interviews! For this iteration, we have interviewed some more veteran students of DCSC and investigated what they are doing now and see if they still apply their S&C knowledge. Some of you might recognise them from past DSA Kalman events.
Name: Gianmarco Disario
Employer: Xomnia
Title: Data Scientist
Year of graduation: 2018
What was the title of your master thesis project and who was your supervisor?
I did my thesis work under the guidance of Jan-Willem van Wingerden (S&C). On the effects of an azimuth offset in the MBC-transformation used by IPC for wind turbine fatigue load reductions.
Is the study material still relevant to your daily occupations?
Yes quite a lot. The control theory aspects not that much, but in the projects I work in there is quite a bit of maths still involved. The basics from SVDs are still quite relevant and currently also working a lot with time series data which even gets me back into some state-space modelling together with quite a bit of probability theory that we used in Filtering & Identification (and Kalman filters).
Do you have any advice for students pursuing a career in your field of work?
Yes, from systems & control you do have a good understanding of many mathematical concepts of algorithms. If you want to stay relevant in data science it important to keep that up. (as well as develop your Python skills, etc)
What does a regular workday look like for you?
A regular workday mostly consists of quite a bit of (Python) programming and thinking of novel ways to optimise business goals with the help of mathematical models.
Have you worked on an interesting project lately?
Yes definitely! I am currently working on a project for Adidas. Due to the corona crisis all/most of the retail stores were closed and there is quite a bit of stock left. Therefore I am working on a project for Adidas.com such that they can optimise the sales by adjusting their pricing strategy on a product level. Sadly I cannot share any further details on it.
Do you have any great memories of studying system and control at the TU Delft you want to share?
The founding of DSA Kalman and the borrels were really fun!
Did you have a favourite or least favourite professor?
Favourite were Jan-Willem & Keviczky. Jan-Willem because of his teaching style, communication and involvement with students as well as a really clear way of explaining Robust & Multivariable. Keviczky because of the structure of his slides and how well he is able to explain all courses that he teaches.
If you have any further questions you want to ask Gianmarco, contact him via
Linkedin
Name: Huib Keemink
Employer: BigData Republic,
Title: Senior Data Scientist
Year of graduation: 2016
What was the title of your master thesis project and who was your supervisor?
I did my thesis at Eneco, supervised by Jan-Willem van Wingerden
The title was: Detecting central heating boiler malfunctions using smart-thermostat data”. Smart (internet connected) thermostats are widespread, but the data they generate is underused. The goal of my research was to prevent cold showers and freezing living rooms by detecting failures in central heating systems before they really manifest themselves.
What are your main responsibilities as a Data Scientist?
My role is to help companies build data-products that create real value. Most companies know they should do something with data, but struggle with moving out of the proof-of-concepts phase where datascience is just expensive maths. This means I write code that can be used in production, coach junior scientists, and also have frequent meetings with the business trying to come up with exciting use cases.
What does a regular workweek look like for you?
My role is to help companies build data-products that create real value. Most companies know they should do something with data, but struggle with moving out of the proof-of-concepts phase where datascience is just expensive maths.
This means I write code that can be used in production, coach junior scientists, and also have frequent meetings with the business trying to come up with exciting use cases.
Do you have any advice for students pursuing a career in your field of work?
My advice would be to focus on where you can add value. Yes, using the latest advances in machine learning research is great fun, but in the end you’re trying to solve a problem. If you really understand the problem, you’ll have a great time solving it!
Did you have a favourite or least favourite professor?
That’s an easy one. Jan-Willem. He was my thesis supervisor and killed it. He was able to point me in just the right direction every now and then, and really helped me in the last few weeks (months?) when I was struggling to get my thesis together in time for my defense.
What was the most difficult S&C course in your experience?
In terms of the exam, Control Theory. Many of the calculations were tedious and required a lot of practice. The course material and the concepts were (in hindsight) not very hard to understand. Unlike the course materials for Nonlinear Control and Filtering and Identification. These were much more difficult. I don’t think I’ve ever really understood them well. Strangely, the exams for both these courses were quite straightforward.
Did you have a favourite S&C course?
Oh yes, there was one course. I believe it was called Control Methods for Robotics. The way it was taught really put all the other things in their place (especially nonlinear control). You know the feeling when it all just “clicks”? That happened during this course for me.
If you have any further questions you want to ask Huib, contact him via
Linkedin