With a background in Information Retrieval and Distributed Database Systems, during the last years I have been interested in reserach projects involving data mining algorithms, big data storage and analysis, machine learning and more recently data pipelines and stream processing.
Working on interdisciplinary projects has always fascinated me. I am curious to learn from and communicate with the professionals of other disciplines. I am specially interested in working on projects on:
- Sustainable Energy
- Media, Arts and Cultural Heritage
- Social Impact
Projects
- TwinDiGrid (Innosuisse, 2022-2023): A Grid Insight Platform composed of a data-driven and real-time digital twin of distribution grids linked to an easy-to-develop-and demonstrate environment. It allows to efficiently generate insights and business values from grid federated and trusted data.
- RST-Control (Hasler Foundation grant, 2021-2022) : Real-time stream processing applied to efficient control of distributed renewable energy sources in smart grids.
- Grid Data Digger (Innosuisse, 2019-2021): An automated Distribution Grid
Operation Assistance Tool using Data-Driven solutions on a big-data Platform. Implementation of the complete grid ingestion platform and the grid data analysis modules using Apache Spark and Spark Streaming. - Dermintel – An AI powered Digital Care Compagnion (2016-2018). Technical supervision leading to the co-funding of the start-up by my Master student.
- DermaQA – Automatic Gerneration of a Dermatology Question Dataset (2018): Automatic detection of similar dermatology-related questions from community-based question-answering forums using IR and deep-learning.
- CrowdStreams (HES-SO grant, 2015-2017): Real-time analysis and monitoring of mobility in the proximity of big events using Apache Spark and ML-LIB machine-learning library.
- Livre Artist (Contract, Bibliothèque National Suisse, 2014-2015): Extraction and analysis of annotations from the complete bibliographic collection possessed by Bibliothèque National Suisse (BNS) in order to characterize artistic works.
- RT-DLP (Contract, Crossing-Tech SA, 2014-2015): Real-Time Content-based Data Loss Prevention (DLP) Technology Feasibility Study. Comparison of Spark, Storm, Hadoop and Hbase for the implementation of a scalable DLP tool.
- SR-DLP (Hasler Foundation grant, 2013-2014): Efficient and Scalable Near Duplicate Detection for Content-based Data Leakage Detection, comparison of four MapReduce algorithms.
- Ef-NDD (RCSO grant, 2012-2013): Efficient Near Duplicate Detection. Improving the efficiency of Near Duplicate Detection algorithms for security audit.
- ClusterSITG (Contract, l’état de Génève, 2012-2013): Automatic clustering and semantic linking of the geographical metadata of the terms used by “Service des Systèmes d’Information et de Géomatique (SSIG) de l’état de Genève”.
- Thesauro (RCSO grant, 2011-2012): Thesaurus Automatic Reorganization. Association rules mining on the RTS (Radio Télévision Swiss Romande) archive to restructure their thesaurus. The result prototype is currently used at the RTS.
- Health Social Media Monitoring (prototype for SwissRe Life & Health R&D, 2011-2012) : A tweet classification prototype to analyse patients, diseases and medications.
- NotreHistoireMobile (Contract, RTS, 2011-2012): Mobile applications on iOS and Android developed following the Mobiwalk platform for the RTS notrehistoire.ch.
- NDD (Contract, Price WaterhouseCoopers, 2010-2011): Design and implementation of Near Duplicate Detection algorithms with high precision. The result was an audit tool tested and exclusively used for one year by PWC.
- Mobilwalk (RCSO grant, 2009-2011): A generic platform to provide multimedia- based services to the users on the move.
- Walking-the-edit (Contract, ECAL, 2008-2009).
In the past, I have been working on audiovisual retrieval for cultural heritage applications
- Video annotation enrichment
- User oriented video querying and browsing
- Audiovisual description standards, MPEG-7 querying