Laurent Candillier French born in 1978 [email protected] (+34)615707752

1

Quick presentation • Engineer in Research & Development in Computer Science - Machine Intelligence • PhD in Computer Science, specialty Machine Learning, Data Mining and Recommender

Systems

• Advanced knowledge also on Information Retrieval and Natural Language Processing • Computer Science technologies : Python, SQL, Java, Apache Spark, Beam, GCP, Elastic-

Search, Scala, PHP...

• Trilingual French, Spanish, English

2

Global cursus • •

• • • • •

3

2017-2018 :

Senior Data Scientist, Ocado Technology, Barcelona, Spain 2013-2016 : Scientic and technical expert in R&D in Computer Science, TokTokTok, Digimind, Target2Sell, TheFamily, Charly.io, SonetIN, Refactor, France Consultants, Overblog, Ebuzzing - Teads 2010-2012 : Director R&D in Computer Science, Nomao - Overblog - Ebuzzing - Teads, Toulouse 2006-2009 : Engineer R&D - post-doctorate (recommender systems), Orange Labs, Lannion 2003-2006 : Engineer R&D - PhD thesis (clustering), Lille University - Company Pertinence, Paris 2001-2002 : Engineer R&D - PhD thesis (proling), Lille University - Company Rosebud, Paris 1996-2001 : Master in Computer Science, Lille - Aston University, Birmingham, England

Research works

The passion for mathematics made me graduate with honors at the end of my secondary school studies. But the playful side of computer science took the lead at Lille University. Discovering Machine Learning during my Master's thesis in the GRAppA laboratory ended up founding my vocation : I want humans and machines to collaborate to both get smarter. In 2001 I started 1

a PhD thesis on Recommender Systems. Collaborative ltering had recently emerged and we shew its complementarity with the content-based proling approaches. Unfortunately, internal issues in the Rosebud company led to the end of this research, and we redirected it to another company, Pertinence, and another domain, Clustering. The angle of our research was to nd a solution that was understandable by humans, but our main contribution stood in the important and open eld of evaluating clustering methods (publication no 1). After my PhD defense I went back to the domain of recommender systems inside Orange Labs. We participated to the Netix prize, did not win the million dollar but obtained very good results with our Collaborative Filtering method optimized by using weighting schemes over similarity measures classically used (publication no 2). Our solution also oered the advantage of producing understandable results to the users. After this post-doctorate I was named director of Research & Development at Nomao - Overblog - Ebuzzing group. Besides my works on sentiment analysis, I co-supervised two PhD thesis, one at Nomao on learning-to-rank applied to geolocalized and personalized search engines, and another one at Overblog on injecting diversity onto recommender systems (publication no 3). Our works on Active Learning applied to the problem of deduplicating items at Nomao also led us to organize a workshop and challenge at the ECML-PKDD'2012 international conference (publication no 4). As scientic and technical expert certied by the Ministry of Higher Education and Research in France, I then followed up young and innovative companies in the elaboration of their R&D projects and the study and development of their intelligent algorithms. In the context of the implementation of the search engine of TokTokTok, we wrote paper n◦ 5 that deals with the interaction between Information Retrieval systems and Natural Language Processing. I am now working at Ocado Technology on automatic fraud detection on Big Data environment (presentation no 6). More details about my research and publications are accessible online at http ://lcandillier.free.fr/CV-eng.php 4

Publications (extract 6/28)

1. [2006] Cascade Evaluation of Clustering Algorithms Laurent Candillier, Isabelle Tellier, Fabien Torre, Olivier Bousquet In Johannes Fürnkranz, Tobias Scheer and Myra Spiliopoulou, editors 17th European Conference on Machine Learning ECML'2006, Berlin, Germany, 18-22 september 2006 Lecture Notes in Computer Science, LNAI 4212, pages 574-581 2. [2008] Designing Specic Weighted Similarity Measures to Improve Collabo-

rative Filtering Systems

Laurent Candillier, Frank Meyer, Françoise Fessant In Petra Perner, editor 8th Industrial Conference on Data Mining ICDM'2008, Leipzig, Germany, 16-18 july 2008 Lecture Notes in Computer Science, LNAI 5077, pages 242-255 3. [2012] Multiple similarities for diversity in recommender systems Laurent Candillier, Max Chevalier, Damien Dudognon, Josiane Mothe In International Journal on Advances in Intelligent Systems, Volume 5, Number 3 & 4 2

4. [2012] Design and Analysis of the Nomao Challenge - Active Learning in the

Real-World

Laurent Candillier, Vincent Lemaire Workshop on Active Learning in Real-world Applications ECML-PKDD'2012, Bristol, UK, 28 september 2012 5. [2016] RI-TAL : le TAL au service de la RI Laurent Candillier, Julien Hénot 13ème Conférence en Recherche d'Information et Applications CORIA'2016, Toulouse, 9-11 march 2016 6. [2018] Machine Learning with Scikit-Learn and Xgboost on Google Cloud

Platform

Laurent Candillier, David Cournapeau, Steve Greenberg Google Cloud Next'2018, San Francisco, 24-26 july 2018

5

Computer Science skills • Universitary teaching : programming in Python, Ada, websites creations, use of internet,

databases, spreadsheets

• Programming languages : Python, Java, Perl, C, C++ • Big Data : Google Cloud Platform, Big Query, Apache Spark, Beam, Scala • Web development : HTML, CSS, PHP, MySQL, PostgreSQL, ElasticSearch

6

Spoken languages • French, spanish, english trilingual • Some notions of german

7

Leisure • Sports, dance, percussions, juggling, reading, writing, chess and cards games, nature,

culture, traveling, ecology, helping people

3