About me

Awards

Invited Talks and Seminars

2024

Anomaly Detection in Time Seires, keynote for AALTD workshop at ECML / PKDD 2024 (September 2024). More info, Slides.

2023

Dimension-wise Class Activation Map for Explaining Multivariate Data Series Classification, Invited talk for GDR CNRS MaDICS Symposium (May 2023)
Detection of Anomalies and Identification of their Precursors in Large Data Series Collections, Invited talk for GDR CNRS MaDICS Symposium (May 2023)
From Anomaly Detection to Precursors Discovery in Time Series, MLMDA Seminar at ENS Paris Saclay, Centre Borelli (January 2023)

2022

Unsupervised Subsequences Anomaly Detection and Precursors Identification for Very Large Time Series, Seminar at VALDA team, INRIA (November 2022)
Automated Anomaly Detection for Very Large Time Series, Invited talk for Forum ORAP : Le PEPR Numpex « Numérique pour l’Exascale » (November 2022)
Detection of Anomalies and Identification of their Precursors in Large Data Series Collections, Seminar at CEDAR team, INRIA, (November 2022)
Unsupervised Detection of Abnormal Subsequences in the Context of Large Data Series Collections, Seminar at BD team, LIP6 (September 2022)
Abnormal Subsequence Detection for Large Collections of Time Series, Seminar at ZENITH team, INRIA (September 2022)
Anomalies Detection and Prediction for Very Large Data Series, Invited talk for INFORSID PhD prize (May 2022)

2021

Easy, Accurate, and Fast Complex Analytics on Big Data Series Collections: Similarity Search and Subsequence Anomaly Detection, Seminar at DEWA R&D (May 2021)

Education

2019-2021: Ph.D. student, University of Paris, EDF R&D
- Analysis of very large multivariate time series using unsupervised learning methods in order to detect anomalies and support predictive maintenance. Collaboration with the National French Electricity Company (EDF).
- Advisors:
  - Prof. Themis Palpanas
  - Emmanuel Remy
  - Mohammed Meftah
2020: Visiting PH.D. student, University of Chicago
- Evaluating a possible extension of subsequence clustering and anomaly detection methods for streaming data series.
- Advisors:
  - Prof. Michael J. Franklin
  - Prof. John Paparrizos
2014-2017: Master and Bachelor degree, Grenoble INP, ENSIMAG
- Engineering and Computer Science School
2016-2017: Exchange Program (Master), Trinity College Dublin, SCSS
- School of Computer Science and Statistics
- (First Honor grade for First and Second Semester)

Experiences

2023: Postdoctoral researcher, Ecole Normale Supérieurs (ENS) Paris Saclay, Centre Borelli
October 2022: Visiting Researcher, European Gravitational Observatory (EGO), the VIRGO collaboration
- Research project:
  - Geo-localization of noises using data series from fiber-acoustic sensors positioned at one of the buildings of the VIRGO interferometer. Collaboration with Dr. Stamatios Gkaitatzis and Prof. Stavros Katsanevas.
2022-2023: Postdoctoral researcher, Université Paris Cité, LIPADE
- Research projects:
  - Benchmarking and large-scale evaluation of anomaly detection methods for univariate time series. Collaboration with Dr. John Paparrizos and Prof. Michael J. Franklin from the University of Chicago.
  - Precursors detection and time series classification for the prediction of le Piton de la fournaise volcano. Collaboration with the Institut de Physique du Globe de Paris (IPGP).
  - Ensembling unsupervised anomaly detection methods with a supervised model selection method. Evaluation on large univariate time series benchmark.
2017-2018: Research engineer, Ecole Polytechnique, LIX
- Research projects:
  - COM4U: Analysis of retail data using unsupervised learning techniques in order to discover causality links between external events and web behaviors. Collaboration with start-up Realytics. Graph mining, url clustering, time series clustering. Implementation in Python.
  - SMARTLAW: Analysis of textual law data using unsupervised techniques in order to estimate the difficulty and the type of trial in the Court of Justice of France. Collaboration with the law faculty of the HEC Paris business school. Text mining, Graph modeling, Graph mining, k-core analysis. Implementation in Python.
- Supervision:
  - Prof. Michalis Vazirgiannis
2017: Research intern, Paris Descartes University, LIPADE
- Development of machine learning algorithms to detect and predict subsequence anomalies in sensor data of electric power plants. Time series similarity/clustering/classification, Matrix Profile (MP). Implementation in C and Python.
- Supervision:
  - Prof. Themis Palpanas
2016: Research intern, IRT Saint-Exupery
- Implementation of an avoidance rover protection system (ARP) in C from an Event-B model. Formal verification of the code using frama-c, why3-based ACSL language.
2016: Fablab project, Grenoble INP
- Design of translator gloves for sign language. Development of a functional prototype using Arduino. Machine learning (KNN-classifier) and data acquisition code (Arduino tools for online data acquisition) done in C++. More info…