Machine learning, deep learning, artificial intelligence—Julien Dinh, Senior Research Lead at...
The banking jobs : Intelligent Automation Engineer
Michele Kidane Mariam is an Intelligent Automation Engineer at BNL in Rome. Here, he tells us about data-driven development processes, the power of today’s computational technology, and much more.
You’ve told us that your role is to redefine how the bank uses data. What does this entail?
First of all, it is the way we look at data. Data is mainly used to store and retrieve things, even today. The way I look at data is to ask how it can enable processes, to enable automation. We are trying to achieve a conceptual shift—looking at data for the value that it has as a data set, not just as a placeholder. We are moving from a point where data is static, used to show the way to something, to data being the king of assets. I always say that the main asset a bank has today is not money, but data.
What implications does this way of approaching data have?
It’s not just a change, it’s a revolution—we’ve heard about the information age since the 80s, but the point to which technology has evolved today now allows us to apply strategies and methodologies where previously there simply wasn’t enough computational capacity. We now have the technological “muscle” to actually do the things conceptualized a long time ago. It’s very exciting. We’ve just come so far. For example, most machine-learning algorithms that we use today were developed in the 70s and 80s, but in those days we didn’t have the technology to execute them. Back then, only the most advanced militaries were able to reach such computational capability. Today I use it at the bank.
Photo : Michele Kidane Mariam
Where do the limits lie today? What challenges does your team face today and how are you working to overcome them?At the moment, the main challenge has to do with how we manage data. One constraint is that we currently only use real data in a production environment, which means that we don’t have a data-driven development process. With a classic IT development logic, you know from the start what the outcome looks like, while a data-driven approach allows the data to shape the solution. To be able to prototype with real data, to play with the data and follow where it leads, this opens up the process to new solutions that we hadn’t thought of at the beginning of a project. For example, one thing we need to do a lot is process real-world identification documents and digitize the data. The first thing you need to do is teach the machine how to read a piece of paper and you need to create a model that is able to identify the target information, however it can be presented. The problem with using fake data is that it’s difficult to reproduce the uncertainty of real-world data, which means we cannot teach the algorithm to deal with real-world circumstances as effectively as we could if we used real data from the beginning of the process.
There is progress to be made there, and we’ve started a process with the IT department to define new roles within the organization that would change the current permissions and access to information and allow these kinds of data-driven approaches. We’re evolving our IT management logic in order to enable this, and that’s a great step forward.
We now have the technological “muscle” to actually do the things conceptualized a long time ago. It’s very exciting.
What do you like most about your job? What inspires you?
In this field, one of the most exciting things is when data reveals something to you that you never thought was worthwhile in the first place. Being able to uncover unknown or unthought-of value out of data that we already have is a great feeling. I like to play with data and see what I can take out of it, transform that data into information. I know this is a strange thing to say, but turning data into information is not as easy as it sounds.
Do you have any projects you’ve worked on that stand out?
The first project I undertook at BNL was great in this respect, because, when I approached the team to implement the solution, I was welcomed and the people couldn’t wait to have something that allowed them to do their jobs better. They said it would allow them to concentrate on the parts of the job that they love, which is a great thing to hear. Now I see this as my objective, to create conditions that allow human beings to work in the best way possible.
Michele Kidane Mariam
Intelligent Automation Engineer Rome
"Problem solving is a key element in what we do. Banking is a highly competitive environment, which means that we have to move at a certain pace, and you have to be able to resolve problems as they present themselves. You also have to understand the everyday business practices of the bank, so you can see opportunities for improvement and automation. Lastly, you need an eagerness to do new things that you haven’t done before."
How would you explain your job to a child?
I would say I use information to determine what to do, and the best way to do it. By collecting a lot of information, we get a better idea of what is going on around us. The more we know, the better the decisions are that we take.
Photo : header ©Christian Hillebrand
Read moreAll news
Ludan Stoecklé became interested in Artificial Intelligence (AI) while a student at INSA Lyon....