BNP Paribas wants 75 people for its new AI unit, based in Paris and Portugal
Once upon a time, the City couldn't get enough of graduates of Nicole Karoui's Masters in Probability and Finance at Paris's at University Paris-VI. Back in the day, Karoui's students were pursued to be derivatives structurers and quantitative risk managers by banks in London. Many of them still occupy senior positions in the City today. However, Karoui retired in 2004 and top French quants say her course has been displaced by contenders more suited to banks' needs of contemporary mathematicians. .
Today's Karouis are Nizar Touzi, professor of applied mathematics at Paris's École Polytechnique and Nicolas Vayatis, director of the, 'Centre de mathématiques et de leurs applications (CMLA)' at the École normale supérieure Paris-Saclay (formerly known as ENS Cachan). The course in question is the masters in applied maths (MVA). Its appeal? It equips France's new best mathematicians with the techniques necessary to work in data science or artificial intelligence.
"The best mathematicians don't want to do derivatives any more." says Edouard D'Archimbaud head of the data and artificial intelligence lab at BNP Paribas in Paris. "The MVA at Cachan is becoming the most famous preparation of a career in machine learning."
D'Archimbaud should know. A graduate from École Polytechnique with an MVA in applied maths, he runs BNP Paribas' data and artificial intelligence (AI) lab. While banks like Citi have comparatively small AI teams, BNP Paribas is ambitious to grow. Right now, the French has 25 people in its lab, spread between Paris and Lisbon in Portugal, ultimately it expects to have 100. D'Archimbaud is therefore hiring - slowly. "We're recruiting around 10 people a year," he says. "We're very careful and only like to hire the right people."
In Paris, those people are highly likely to be students of Touzi and Vayatis. Thanks to their endeavours, French graduates looks likely to retain their position at the helm of quantitative finance as the sector embraces machine learning. "There's a saying that the English are good at law, the Americans are good at history and the French are good at maths," jokes D'Archimbaud's colleague.
In Portugal, D'Archimbaud's unit is more likely to hire from the Universities of Lisbon or Porto. While banks like Morgan Stanley are building quant units in Hungary, D'Archimbaud says Portugal suits BNP Paribas best: "We find great profiles there, it's close to Paris, and it's cheap."
Speaking at this week's AI summit in London, D'Archimbaud said his unit is developing new interfaces between clients at the bank. So far, it's worked on a translation program for finance which he says is more accurate than Google's own.