Senior Project Manager

• May 2023 - Present

Project Manager

• Oct 2022 - May 2023

Head of Facility

• Jan 2021 - Oct 2022

Head of the Digital Health Facility (DHF). In charge of managing the DHF’s personnel, and projects.

Senior Specialist

• May 2019 - Dec 2020

I was in charge of developing machine learning (ML) algorithms for smart diagnostics. Current projects involve both academic and industrial partners, and aim at predicting clinical outcomes by combining machine learning algorithms with omics data, medical images (mostly using deep learning), and wearable sensors (e.g., EEG, ECG, IMU). Moreover, I am involved in writing proposals (e.g., InnoSuisse and private foundations) to fund the research work of the institute. Furthermore, I am responsible for the IT of the SICHH and the Digital Health computation unit. Show less

Scientist

• Dec 2014 - Dec 2018

I am a scientist at the Atmospheric Particle Research Laboratory (APRL), and I develop and apply machine learning algorithms to interpret multivariate infra-red (IR) spectra of ambient aerosol collected on Teflon filters. I focus on the organic and inorganic mass of the aerosol. The methods we develop aim at (i) removing substrate interferences (Teflon fingerprint) from the interested signal (baseline correction); (ii) quantification of functional groups (alkane CH, ammonium, carbonyl, ester, ketone, carboxylic acid, alcohol OH, amine, organic nitrate, organic sulfate, and carboxylate); (iii) multivariate calibration of IR spectra of ambient or laboratory generated aerosols. The work involves the utilization of the measuring device, data pre-processing, algorithm developing, data evaluation, interpretation, and analysis reporting. I also developed and maintain the AIRSpec, a graphical user interface to make the developed algorithms available to users without a computer science background. The interface is based on the Shiny R library (http://airspec.epfl.ch/). I am also a teaching assistant in the course Air pollution and climate change, in which I help the students in theoretical and practical tasks. In particular, I help students in analyzing and interpreting air pollution datasets using R. Goals: I deepened my knowledge about the organic and inorganic aerosols. I gained experience in characterizing aerosols using Infra-red spectroscopy and learned how to deal with multivariate data analysis. I mastered my knowledge in coding with R (included shiny), Python and Matlab. I also updated my knowledge of ML algorithms. All these achievements result in high-quality publications in peer-reviewed international journals and conferences. I also improved my experience in teaching, helping and reporting results. Show less

Freelance Consultant

• Jan 2013 - Nov 2014

I was a Consultant/Freelancer (self-employed) in the domain of sensor networks, environmental monitoring and mapping, and graphical user interface developer. In this period: - I worked 30% of the time to finalize my Ph.D. thesis: (i) I repeated some simulation experiments (around 1000 trials), and I rewrote three chapters based on the new results; (ii) I added 3 appendix chapters. - I worked 20% of the time in collaborations with research institutes for scientific articles for Elsevier journals: we finalized three manuscripts. - I worked 20% of the time in attending courses (Coursera platform) to improve my scientific writing and my coding skills. - I worked 10% of the time preparing an EU proposal for funding in the field of sensor networks for environmental monitoring. -I worked 10% of the time as graphical user interface developer. Show less

Scientific Researcher

• Feb 2011 - Dec 2012

I was a Researcher at Flemish Institute for Technological Research (VITO) Environmental risk and health unit. I worked mainly on two projects. 1) The IDEA had the primary interest of valorization for the benefit of society. The project focused on the development of local, intelligent measurement networks for noise, UFP (ultrafine particulate matter), and gas pollutants, which demand a high spatial and temporal resolution. We developed cost-effective nodes using commercially available (low cost) gas, temperature and humidity sensors. In detail, I compensated for the loss of measurement data quality of the array of low-cost sensors by applying machine learning techniques (ML): (i) signal pre-processing (e.g., filtering, baseline correction); (ii) feature extraction (e.g., DFT, DWT); (iii) dimensionality reduction (PCA, LDA), (iv) supervised learning model: regression (Gaussian process), classification (SVM, RVM). Moreover, I performed sensor network calibration and anomalies detection, prediction of UFP from commonly measured pollutants (NOx, CO, O3). We also developed a mobile platform to carry pollution measurements. I designed and performed, in collaboration with colleagues, measurement campaigns and strategies to test and evaluate the models. 2) The EveryAware (EU/FP7) project aimed at raising the environmental awareness of people that in turn will trigger more sensible behaviors. The project focused on the development of hardware and software to collect pollution data at affordable costs and in the aggregation of a large number of records collected from different users. My tasks: - Contribution in the design of the EveryAware platform, consisting of a gas sensor box, a smartphone, and the back-end system: (i) I designed the sensor box; (ii) I selected the sensors and hardware in collaboration with colleagues; (iii) model and software for gas sensor array calibration based on ML; (iv) model and software for spatial-temporal pollution distribution maps. Show less

Researcher Phd Candidate

• Jun 2007 - Sep 2010

Main tasks: 1) Research in gas distribution modeling for mobile robots applications. The main realizations were: (i) set up of a prototype robot equipped with an ultrasonic anemometer, electronic nose, and laser scan; (ii) software development robotic applications; (iii) Machine learning models for 2D and 3D gas distribution maps (Kernel DM+V/W algorithm). 2) I worked in the project: DustBot - Networked and Cooperating Robots for Urban Hygiene. The DustBot aimed at designing, developing, testing and demonstrating a system for improving the management of urban hygiene based on a network of autonomous and cooperating robots, embedded in an Ambient Intelligence infrastructure. The project focused on the development of two types of robots: DustCart and DustClean. DustCart is a two-wheeled robot, designed to provide an automatic door-to-door garbage collection; DustClean is a four wheels mobile base equipped with tools for sweeping and collecting garbage from the ground. We developed Ambient Monitoring Module (AMM) and integrated into the whole system. The key idea was to feature robots that are primarily used for something else with the capability of pollution monitoring and detection of air pollutants. The main realizations were: Contribution in the design of the AMM module: (i) Design of the electronic nose; (ii) Software for Web server; (iii)Software for spatial-temporal pollution distribution maps (Kernel DM+V algorithm); (iV) Testing and Demonstrating. 3) I taught for three years at Örebro University three laboratory courses, and I also supervised a master thesis. Goals: The main achievement was to obtain the Ph.D. degree in computer science. I laid the foundations of my research career. I learned robotics, machine learning, spatial statistics, how to code, conduct research, writing articles and preparing good presentations. I got my experience in coding in Matlab and in c++. All these achievements result in high-quality publications and conference presentations. Show less

Research Assistant

• Jan 2004 - May 2007

I was an assistant researcher at Budapest University of Technology and Economics in the Department of Electron Devices. I mainly focused my research on the characterization of electron device surfaces. My main tasks were: 1) Characterization of the surface and bulk properties of different devices using Kelvin probe (vibrating capacitors) with additional high-intensity illumination and SPV (surface photovoltage): (i) photovoltaic devices based on silicon and thin films (CIGS) on a glass substrate; (ii) gas sensors 2) Utilization of the Kelvin probe with additional high-intensity illumination to measure the sensitivity and selectivity of gas sensors. 3) Clean room work: (i) fabrication of gas sensors based on porous silicon doped with noble metals; (ii) deposition of thin film layers on silicon, glass, and ceramic substrates; (iii) fabrication of silicon solar cells. Goal: Results presented in international conferences. Show less