Predictive Modelling & Data Science at Citi
I studied Mathematics in Italy, the UK and the US, and I did a post-doc in France. I then worked in satellite image analysis where I started to focus on Machine and Deep Learning in order to automatically classify and categorise satellite imagery. I co-authored a book on Deep Learning and I am currently working at Citi on the US Consumer Analytics team.
Having a varied background, with a strong foundation in Mathematics and especially high dimensional Geometry, is very important for becoming an expert in Machine Learning and Deep Learning. Data science and analytics problems require a deep understanding beyond a simple application of algorithms. By defining the right metric and coordinate system, one can then really understand, and solve, analytics problems.
How did you get to become an expert in your key topics?
Studying theoretical Mathematics, not applied, enabled me to look for the basic underlying concepts behind any theory. When I started to work on computers and learn computer languages, I was able to design programming code that was inherently cohesive and well-structured. Deep Learning and Neural Nets, and Machine Learning in general, was the natural evolution of my career because I could use Mathematics to create computer algorithms that extracted key concepts inductively from the data itself. I think of data, any data, as a subspace in some multidimensional space and I then look for the function that most closely outlines its shape. My approach is to implement algorithms that describe this structure and look for solutions to data problems the same way I look for solutions to mathematical problems.
What sub-topics are you most passionate about?
I am passionate about Deep Learning and Neural Networks even though they can still be limited in their applications. At Citi, I am passionate about creating the most useful experience for our customers by delivering products that are personalised to each individual’s financial situation to help them reach their financial goals.
Who influences you within these topics?
Yann LeCun is one of the most influential researchers in Artificial Intelligence, but I also appreciate Gary Marcus’s efforts to show the innate limitations of Neural Networks. They represent somewhat alternative approaches to Artificial Intelligence. While most researchers situate themselves on one side or the other, it is important to understand both perspectives behind the complex question of “what is intelligence.”
What challenges are brands facing in this space?
It really depends on the type of brand and company. A major concern around Artificial Intelligence is bias. Machine learning algorithms should not repeat any bias already present in the data. If an algorithm shows bias, then you can try to improve the algorithm through explainability. Explainability lets you understand how the algorithm reached a conclusion and provides an opportunity to eliminate any issues caused by a poor selection of features and/or to balance the data to remove the sample bias. Every company, regardless of the business they are in, needs to work to eliminate any bias in their process to ensure their products and services are equitable and fair for all their customers.
What do you think the future holds in this space?
Personalisation is going to be key. Customers want a company that understands them and provides personalised services. In financial services, FinTech companies have been leading the way but, major banks are also focusing on personalisation to better serve their customers while complying with regulation. Personalisation while protecting consumers’ privacy will be an important trend for companies, regardless of industry.
What are your passions outside of work?
I am interested in several topics, ranging from neurology to quantum physics, history and the arts. I believe knowledge is the product of synthesis, not specialisation. I also love to write short stories and novels although I have not yet published any.
What would be the best way for a brand to contact you?
While you’re here, why not signup to our B2B influencer marketplace, MyOnalytica? MyOnalytica is the world’s largest B2B influencer marketplace. Influencers can sign up for free and create their own profiles which will be visible to Onalytica clients which include many of the world’s largest brands. Sign up now to showcase your expertise, influence & how you would like to partner with brands.