Personal Picture

Daniele Gadler

daniele.gadler at yahoo.it




💡~/About

You may download the latest version of my CV at this URL.

I am a trilingual data scientist fluent in English, German and Italian, passionate about managing data and productionizing machine learning models and data science solutions. I hold a MSc. in Computer Science and Networking (cum laude) from the Scuola Superiore Sant’Anna and the University of Pisa, from which I graduated in 2019.

Data scientist experience: From March 2021 until December 2023, I have been working as a data scientist at One Data GmbH in Frankfurt am Main, Germany on the optimization of the supply chain for the production of medicines. Before that, I worked at Südtiroler Informatik AG and developed models to predict COVID insurgence in patients based on their symptons.

Research experience: At the National Research Institute of Pisa, I worked on research for Recurrent Neural Network quantization in IoT devices. At the Free University of Bolzano-Bozen, I investigated the usage of Hidden Markov Models for modelling system usage in complex systems, and at the Technische Universität Kaiserslautern studied techniques to identity Wi-Fi devices in networks based on packets’ electromagnetic fingerprint.

👨‍🎓~/Education

🚀~/Work Experience

🏆~/Awards and Scholarships

📄~/Certificates and Habilitations

📝~/Publications

📢~/Talks

㊗️~/Languages


Native
Proficient - fluent
Advanced - fluent
Discrete

💻 ~/Projects

SIAG - Machine Learning for Traffic Forecast (2019 - 2020)

Machine Learning prototype aimed at predicting road traffic trends in South Tyrol based on IoT and weather sensor data via neural networks, random forest and decision tree models. It uses a Flask frontend for data visualization on a Google Map.

WeAgainstVirus Hackathon - PharmaQ Computer Vision (2020)

Web portal for pharmacy queue estimation during the Covid-19 pandemic. Grab your phone, take a picture, upload it on the website, and the number of people waiting in queue is detected and reported on an interactive map! Based on Flask, a server DB, a faces obfuscator and Nanonetes' API for people detection.



OpenMined - PySyft project (2019 - 2021)

Collaborator to the PySyft project of the OpenMined community with code fixes for the training of Recurrent Neural Networks, Long-Short Term Memory Models, Gated Recurrent Units and optimizations for low-power IoT devices such as Raspberry PIs.

SPADE - Sensing, Processing and Analyzing Data in an Embedded IoT Device (2019)

Implementation of an ML-powered edge computing system for the predictive maintenance of machines, based on a Beagleboard green and an accelerometer. It models the sensors' vibrations using a PCA algorithm and a Neural Network autoencoder to detect abnormal machine behaviour and reports the state on a Grafana UI.



Automatic Hidden Markov Model Generation (2016 - 2019)

Scripts to generate Hidden Markov Models automatically and manually, based on the paper " Mining Logs to Model the Use of a System " implemented in Java Hadoop and R.

Sobel Filter on CUDA, C and OpenCV (2019)

Native C, CUDA, OpenCV versions of the Sobel Filter, an algorithm computing the countour of an image based on the partial derivatives of an image over the X and Y axes.



Persistent World Scripts for Warband (2014 - 2018)

Set of scripts for the "Persistent World" and "Persistent Kingdoms" mod of Mount & Blade: Warband that add roleplay and persistent functionalities to the game, connected to a MySQL Database and a PHP backend.

Data Mining (2018-2019)

Application of data mining techniques such as clustering, supervised classification and association rules to the task of banking customer default prediction in Python. "Lost in Data" team in the 2018-2019 Kaggle challenge.



Dot Matrix Printer (2014-2015)

Hand-made printer using felt tip pens as ink, created based on an Arduino and a Raspberry PI. The system is fully accessible via a web-based interface for uploading images to be printed.

JSnake (2015)

The famous Nokia phone game "Snake" implemented with the Java 8 programming language using, among others, multithreading and generics.



💼 ~/Lecture Notes


Lecture Notes taken during the Master's in Computer Science and Networking 2017-2019 in Pisa and during my Bachelor. All the material released is intended for didactic use only.

Academic Year 2018-2019

Academic Year 2017-2018

Academic Year 2016-2017 - Technische Universität Kaiserslautern