Giorgio Ruffa Email & Phone Number

Stoccolma, Stoccolma, Svezia

I work on RayIntelligence, the next-generation Oncology Analytics Platform.

Data Engineer, Machine Learning department.

Stoccolma, Svezia

Master Thesis Student in the Machine Learning Department

Madrid, Spagna

As a data science intern, I mostly worked with our patented Visual Trending Technology; a powerful fingerprint-based system able to generate recommendations considering solely visual resemblance. I focused mainly on the applications concerning the retail market.More in detail:• Analysis of the performance of the product with a purely data-based approach by leveraging aggregated data provided by the customer. The net result was the development of a single scalar metric able to classify the performance of each recommendation. I proposed numerous strategies and modifications in order to achieve better results.• Creation of Dashboards and Visualizations of key performance metrics for our products. The dashboards were made using Kibana and Elasticsearch and were used directly by all the stakeholders in the company.• Development of a pricing model to optimize recommendations results taking into account items' prices.• Network and cloud infrastructure troubleshooting and debugging.Technology stack:python3, pandas, Elasticsearch, Kibana, AWS, bash, linux.

Milano, Italia

Main projects:• Studied the behavior of existing sound matching algorithms in constrainedscenarios in order to improve their performance. Used C++ to implementthe final algorithm and python+pandas+bash to simulate its behavior on awide variety of input samples. Accuracy was evaluated in function ofdifferent variables to predict the algorithm’s suitability to productionenvironments. The code proved to be very useful also in detecting devices’malfunctioning and incorrect behaviors.• Developed Linux Embedded firmware for UHF RFID high performancereaders (Impinj SpeedwayTMseries). LLRP protocol was used to instruct thereader to write, read and lock UHF tags, always trough a non-blockingasynchronous implementation. I was responsible for the design,implementation, tuning and automated functional testing of the product,along with documentation drafting. Stress tests and long run tests wereperformed to ensure high performance and proper memory management.• Written two Android libraries and a sample app to implement ourproprietary encrypted BT4 mesh protocol for home automation. Worked instrict collaboration with the project manager and the hardware engineer tocomplete the prototyping on a very tight schedule.• Implemented various ETLs to collect data from external services (Vodafoneand Telefonica SOAP API) as well as internal sources (logs, databases, hwmetrics) related to our IoT infrastructure. Created profiling metrics andmonitoring tools fully integrated with Icinga (on premise) and Cloudwatch(AWS boto3).

Milano, Italia

During a series of lectures on c++ programming at the department of Physics of the University of Milan, where I was both a student and a system administrator, I had the chance to meet Stefano Magni, CEO of Akerue srl. After some talks and good amount of hands-on/real-life problem solving, I was offered a part-time position as Junior Software Developer. This fortunate encounter marked the beginning of a rather challenging, but nevertheless incredibly stimulating period, where I had the chance to grow professionally while completing my studies in Physics. It wasn't easy but it was definitely worth it.Some of my early achievements were:- Developed client library for our internal SMS service. Language: c++. Framework: Qt, full TCP/IP stack and asynchronous event system. The library is still used in mission critical environments for alarm notifications.- Developed a Qt native library to handle json configuration files. The library quickly became a fundamental component of all our c++/Qt projects, both server-side and client-side.- Successfully implemented a batch monitoring system for vast pools of devices (10K+), using c++/Qt and MySQL. I actively participated to the design process. The produced metrics are used to identify malfunctions or incorrect installations. The system is still in production and is actively used by two important costumers.

Milano, Italia

I was in charge of the maintenance, monitoring and upgrade of two diskless-root Debian-linux clusters.The first cluster was composed by 60 workstations used for didactic purposes (courses, lectures, training) by students and professors.The second cluster was a computational farm of 200+ CPU and 6 GPU used by the theoretical physics department of the university of Milan.We maintained a variety of servers: the NFS server, log servers, a firewall/gateway, a webserver within a DMZ implemented using VLANs, a mail server, various monitoring tools (Nagios, Zabbix,...) along with ISC-DHCP, DNS and backup servers.I learned to use various networking tools like tcpdump, wireshark, SNMP libraries, that I still use proficiently nowadays.I have personally built the farm's log server using rsyslog and a MySQL for persistent storage. The disks were configured using a software RAID 5 (mdadm).To fulfill the needs for a better monitoring system to keep track of lab's printers usage, I wrote a custom tool in python to interrogate the CUSP server and collect useful information on the number of pages printed by each student, including qouta management. Writing and maintaining bash scripts was obviously a big part of the job. I have also installed and configured an HTCondor pool of the didactic cluster. HTCondor is an High Throughput Computation system aimed to harvest as many unused computational resources as possible. This pool was used by many students to run simulations for their thesis, some of them lasted for weeks.I was incredibly glad that part of my role included the possibility to teach to less experienced users what I have learned. I was also very lucky to have the opportunity to work with amazing people that were willing to share their knowledge in an open and trustworthy environment.

Master Universitario, Data Science

Second year of a Master's Degree program with a major in Distributed Systems and Data Mining for Big Data and a minor in Entrepreneurship..

Master Universitario, Data Science

First year of a Master's Degree program with a major in Distributed Systems and Data Mining for Big Data and a minor in Entrepreneurship..

Bachelor Of Science (Bsc), Fisica

Activities and Societies: Laboratorio di Calcolo e MultimediaCoursework: • Mathematical Analysis • Linear Algebra • Quantum Physics •.