Yannis is a Machine Learning Engineer. He joined the team in November 2017 to work on machine learning assignments such as language understanding, semantic analysis and chatbots. In this article he’ll be telling us about what it was like to join iAdvize: what the recruitment process involved, what the first few days were like, how things are going 3 months later… how he’s survived so far!
So Yannis! Your kind colleague Iori named you for this interview. How do you feel about that?
Hi Miranda! I’m glad you talk about this, we have a special connection with Iori since we arrived at iAdvize on exactly the same day and we both work in the R&D. He’s a great colleague and even greater individual – but I’ll take revenge for this ;)
You trained in France and Denmark and got a double diploma in machine learning. Can you define machine learning for us common mortals?
Ahah yes! Machine learning is a subset of artificial intelligence, which is still very much in its infancy. Unfortunately, both terms are suitcase words, and it contributes to its hype and misunderstanding. To be honest not knowing what exactly machine learning is seems legit.
With machine learning you can automatise many repetitive and tedious tasks that you could, in a favorable environment, accomplish within a few milliseconds. Some examples are speech recognition or image classification. There are also more complex tasks for instance with language understanding and conversation modelling.
How does this work? Machine learning algorithms don’t tell the computer line by line what to learn. This would be very long and non-exhaustive. Instead, they tell the computer how to learn.
Let’s take an example: you want to look in Iori’s library for all images of cats. You don’t want to code that by hand for two reasons: first there are simply too many pictures of cats in Iori’s lib, you don’t want to hand code all the details of all their faces and coat so that the computer can recognise them – second, you would likely miss some features.
So you decide to build a machine learning model that will learn by experience all those features by looking at thousands of photos of cats that you downloaded from the internet. You can see this as a big organ (the model) where the length of the pipes (the parameters of the model) are changed so that a combination of notes (the pixels of the image) produces the right sound (the prediction of whether there is a cat in the picture or not).
How did you hear about iAdvize? Why did you decide to apply?
I heard about iAdvize during a machine learning meetup, in Nantes. After the event, Thomas from the AI team contacted me and we discussed about the challenges the company is willing to solve using machine learning, and it was a fit.
In my previous job I was a ML researcher in a Danish startup, working on image recognition and object detection. I was the only one working on the artificial intelligence deliverables of the company. For my next job, I wanted to find a position where I could be in a team with a strong ML background, where I could still do strong research and development. This is one of the main reasons why I decided to apply.
What were the different stages in the interview process? Did one encounter in particular convince you to join the team?
The interview process was flawless and lasted somewhere between 2 or 3 weeks. I first had an informal lunch interview with Thomas. He gave me a better sense of the projects iAdvize is working on and what my missions in the company would be, this was also a way to check if there was a cultural fit between the company, me and the team I would be working with. I then had a call with Goretti from the HR team. Meanwhile I completed a technical test and I went to iAdvize for an interview with Fred, VP Engineering, who is my manager now. I appreciated the transparency from iAdvize and the mutual confidence that was built upon it.
What happened during your first week here? Which aspects of the onboarding process really made you feel at home?
During the first week I mainly forgot a lot of names :) But I think everybody does!
There was also many other things that I did during the first week together with the other new employees. It was very intense – iAdvize puts an emphasis on making you feel very welcomed and easily integrated. When you arrive you need to have a really good understanding of the product and a clear overview of the whole company with the different teams and their missions: there are mountains of information but it’s worth it and that’s how you get into it!
Any suggestions for what we could do to make it better?
If I need to find something I think I would have liked to have more newbies meet oldies, that’s a great way to meet people from other teams. It’s also nice to have a better understanding of other’s job in the company.
What are you looking forward to in 2018 at iAdvize?
I’m not really looking forward something in particular at iAdvize – except being successful with a successful team and having a lot of fun. I decided to apply for the position I am in because language understanding is what I like the most in machine learning and because it is a very challenging topic.
And to end this interview, a piece of music or track we should all listen to now?
As some may know, I like to pretend I’m a DJ. I often end my sets with this song: https://www.youtube.com/watch?v=UYj8nGus-c0