About Me
Hi, I am Stefano Huber and I am a PhD student in Theoretical Computer Science. I studied at EPFL and Unitn, I am passionate of algorithms and theoretical CS.
When I am not working, I like practicing sports: tennis, basketball, riding my mountain bike and hiking are my favourite!
Education
I am a PhD student at IDSIA USI-SUPSI. Previously I had my master's in CS at EPFL and my bachelor's in CS at the University of Trento. Before university, I have attended the Scientific High School Galilei Trento, where I have developed my passion for computer science taking part in multiple robotics and coding competitions.
IDSIA USI-SUPSI: PhD in theoretical computer science (Sep 2023 -> Sep 2027 (estimated))
My research focuses on the adversarial complexity of falsifying heuristics. This is a topic of deep interest in Complexity Theory: if, for an adversary, finding counter-examples of a heuristic is "hard", then the heuristic is working in practice like an optimal algorithm.
I am working under the supervision of prof. Palmo Monaldo Mastrolilli.
EPFL Master’s in computer science (Sep 2019 -> Jul 2022)
The courses I took focused on algorithms, probability and machine learning. The courses I am most fond of are (in no particular order):
- Advanced Algorithms: focused on matroids, linear programming, probabilistic and approximated algorithms.
- Topics in Theoretical CS: all about crazy spectral graph theory.
- Advanced Probability and Applications: where we covered, among others, martingales, convergence of random variables, random vectors.
- Probabilistic Methods in Combinatorics: where we used the probabilistic methods to prove many unexpected results.
- Artificial Neural Networks: where we learned the basis of reinforcement learning (and how to play pong).
- Applied Data Analysis: for friends ADA, where we wrote some code. A lot of code.
I am co-author of a paper about entity-matching, written during my semester project (autumn 2020) at dlab. The paper has been published at VLDB.
I wrote my thesis on Spectral Clustering in Hypergraphs under the supervision of Michael Kapralov.
University of Trento, Bachelor’s in computer science (Sep 2016 -> Jul 2019)
I graduated with 110/110 with honour, my theses focused on link prediction algorithms and was written under the supervision of professor Alberto Montresor.
The best course I have taken is definitely algorithms and data structures, where my group and I (fatine_ricorsine Italian for recursion fairies) managed to be the top team during the projects!
Scientific High School Galileo Galilei Trento (Sep 2011 -> Jul 2016)
I graduated with 100/100. But, out of all the subjects that I have studied, what made this period so special were the competitions that we did out of scope of our classes: with my robotics team Rescue Maze Galilei, we achieved 1st and 2nd place in the Robocup Italy consecutively in 2015 ans 2016 (for the category Rescue Maze), and got 3rd in the 2015 international Robocup competition in Hefej, China.
Working Experience
While studying at university, I managed to gain working experiences working part-time or doing internships for some companies:
Verity AG (Jan 2023 -> Aug 2023)
I did an internship at Verity AG in Zurich, where I managed the coexistence of drones and people working together in the same warehouse.
QBOID (Feb 2021 -> Feb 2022)
I worked for the Californian startup QBOID, where I solved computer vision tasks using OpenCV.
SchoolCalendar (Feb 2020 -> Present)
I am one of the main contributors of the open-source web application SchoolCalendar: this tool allows school headmasters to manage the professors’ and courses’ workload: as crazy as it may seem, generating a proper timetable is an NP complete problem (so full credits to whom is able to do it on paper)! At the moment two schools are adopting it (CFP Tesero and IFPA Levico), both in Trentino region, with promising results.
SpazioDati (Feb 2019 -> Jun 2019)
I did an internship with SpazioDati, under the supervision of Alessio Guerrieri. There I explored link prediction algorithms, applying them to their vast temporal datasets of the application atoka, a popular “social network” of Italian companies.
Motorialab (Sep 2017 -> Jul 2019)
I have been working part-time for two years at Motorialab as a backend developer using the framework Django. Their biggest product is Safe, a platform where ski-areas administrators can register the accidents that happen on ski slopes. Here some pictures of the team having fun :)
Carluccio’s (Jul-Aug 2017)
During the summer break of 2017, after having finished my exams, I worked as kitchen porter at the London Carluccio’s restaurant. Although it didn’t help me much in developing my coding skills, it certainly improved my English!
Awards
Since high school, I have been extremely fond of coding competitions: they are a great way to improve my skills of working under pressure, they are extremely challenging and, more importantly, they are great fun!
Some awards that I achieved during the years:
- Winner of the 2020 edition of Bending Spoons’ First Ascent (online, guess the year), after having solved multiple challenges of logics and coding.
- Finalist of 2019 Bending Spoons’ (24th overall) CodeFlows Milan, algorithms competition divided in multiple rounds.
- Winner of Vodafone Tobi Hack Trento, a 24h non-stop hackathon where we helped Vodafone’s bot Tobi to achieve better performance.
- 2nd place in Cy4Games in Rome, an attack-defence CTF.
- 1st Italian university team at 2018 Google Hash Code (RecursionFairies), a competition where we had to develop an approximated algorithm to assign in an profitable way taxis to customers.
- Bronze Methal at the Italian Olympics in Informatics OII 2015, an algorithm competition.
- 3rd place in 2015 Robocup Jr Hefei in the category Rescue Maze. In order to participate, we built a robot that managed to go through a maze without getting lost, rescuing victims. Even though we had a large economic disadvantage (some of our competitor’s wheels where more expensive than our entire robot), we still achieved 3rd place! Here a video of our best run.
- 1st Place in 2015 Robocup Jr Italy: we were the first team to win in the category Rescue Maze, using open hardware Arduino, as opposed to the closed hardware Lego Mindstorm.