François-Xavier joined Margo on leaving the Ecole des Mines de Saint-Etienne. He tells us how his career has developed through training and missions.

Can you tell us about your career at Margo?

I started at Margo after leaving Mines de Saint-Etienne. To perfect my skills in information technology and finance, I had the chance to join the C# Academy at Margo.

Following 12 weeks of intensive training, I joined a large financial and investment bank for several months on a project in C# dealing with counterparty risk on primary materials.

My second project was more practically oriented – I had to spot fluctuations in Greek calculations using machine learning.

After a year, I applied to HR to work on data analysis, and Margo supported my request.  I quickly got several mission proposals and was able to choose to continue my career in London and move into automated trading systems. Since then I’ve been part of a research team and I collect, clean up and make available all the data the teams need to set up and run automation models.

Can you tell us more about the C# Academy and your very first post?

I spent 3 months at the C# Academy. We started by learning the language basics, then we saw how to use it more efficiently, for example by reducing memory fragmentation.

Then we went on to more specific subjects with reflection and parallel calculation issues. Throughout the training, we had the use of code handling and analysis tools which I was then able to draw on during my missions.

I put these skills to use from my first job, where one of the main areas was collecting and collating information on scripted deals to send to the risk team for pricing. That was a great introduction to finance for me!

Now let’s talk more specifically about your new challenge in London and the technical constraints involved.

I’m part of a team in a finance and investment bank specialising in algorithmic trading. I work in a quantitative research laboratory which provides data to the Quants. They use it to work out and validate trading automation strategies. It’s a very intellectually challenging environment.

As for technical constraints, several gigabytes of data are currently generated each day. But when you’re talking about 15 or 20 years of historic data to deal with, things get more complicated! So, first, there are problems with storage and then with distribution of the data to the servers to run the calculations.

How is the finance field interesting data-wise?

In automated trading, especially in respect of low and mid-frequency strategies, historic data is crucial for setting up models, which then let you test and validate the strategies. But for that you need the maximum information in a data sample of a size that’s reasonable in relation to the computing power available. So you have to manage to select the important data without distorting the information.

To conclude, we hear from his Business Manager:

“François-Xavier works in an international environment with one of our most demanding clients. He works on very complex data problems in a team of experts on the subject. A chance to develop his potential!”