Internships
HiPeRT lab is offering many and diverse opportunities for students to join the group line of research and technology transfer with industrial partners. It is an exceptional chance to both be challenged and grow as a working computer scientist and engineer, and to make contact with relevant companies.
Prospective students are invited to find topic of interest in the subjects list for theses, projects and internships and to contact the proponent tutor. More general informations have been collected in the frequently asked questions on theses and internships.
Internship list
| Title | Description | Tutor | External partner/project | Subjects | Suitable for |
|---|---|---|---|---|---|
|
Design of an runtime library for dynamically reconfigurable server infrastructures |
The thesis focuses on porting on a highly-reconfigurable server-like machine a runtime library for supporting parallel programming models (e.g., OpenMP, CUDA, OpenCL),and then modifying it so that it can reconfigure the machine on-the-fly to cope with dynamic workload changes. If he wants, the candidate might spend some time at University of Ferrara, which is partner in the project |
Reconfigurable servers, cloud computing, parallel programming |
Bachelor/Master |
||
|
Implementation of an autonomous driving vehicle in 1/10 scale with embedded parallel computers (GPGPUs, FPGAs…)
|
Implement, test and optimize an advanced software stack for autonomous vehicles in scale 1/10, using either GPU accelerators (CUDA), or reconfigurable platforms (FPGAs) |
Parallel computing, autonomous driving |
Bachelor/Master |
||
|
GPU-Accelerated ADAS: models, tools and simulation |
Acquire knowledge about automotive safety standards and modeling tools (i.e: AMALTHEA) and extend these existing tools so to account for modern GPU-accelerated embedded boards and related APIs (CUDA) |
Bosch (DE) |
Real-Time systems, Safety critical applications, HW architectures. |
Bachelor/Master |
|
|
Technical measures for GDPR compliance |
The purpose of this research is to investigate IT technical measures in order to comply the practice of companies, operating in mechanical sector, with the GDPR, including through an internship. |
Fondazione Marco Biagi and/or business with which it has an agreement |
GDPR, privacy, Technical measures |
Bachelor/Master |
|
|
Jailhouse hypervisor on Nvidia |
Jailhouse is a Linux based open-source hypervisor for embedded and safety-critical applications designed by Siemens. Highly responsive, easy to use and extend, it has been successfully chosen at for several research initiatives aiming at improving its real-time capabilities.
The goal of this project is to further extend Jailhouse support for the latest high-end Nvidia heterogenous systems on chips, e.g. Xavier. The activities range from simple tasks like providing basic functionality on a new board, or profiling a specific functionality, to more complex tasks like designing and tailoring new Jailhouse features.
The activities may include direct collaboration with Nvidia and may reach original scientific contributions.
The worker will join a local team of developers, interns and young researchers working on similar topics or systems. |
Nvidia and other confidential parties |
Real-Time systems, embedded systems, operating systems |
Bachelor/Master |
|
|
Real-time for Jailhouse hypervisor |
Jailhouse is a Linux based open-source hypervisor for embedded and safety-critical applications designed by Siemens. Highly responsive, easy to use and extend, it has been successfully chosen at for several research initiatives aiming at improving its real-time capabilities.
The goal of this project is to improve Jailhouse core features, for instance by: reducing the coupling with Linux, lower the boot time; implementing research-level memory management technologies, such as advanced co-scheduling techniques; extending support to the latest heterogenous platforms.
The activities may include direct collaboration with industrial partners from the automotive or chip designer domains, and may reach original scientific contributions.
The worker will join a local team of developers, interns and young researchers working on similar topics or systems. |
Confidential parties |
Real-Time systems, embedded systems, operating systems |
Bachelor/Master |
|
|
TSN for Jailhouse hypervisor |
Jailhouse is a Linux based open-source hypervisor for embedded and safety-critical applications designed by Siemens. Highly responsive, easy to use and extend, it has been successfully chosen at for several research initiatives aiming at improving its real-time capabilities.
The goal of this project is integrate TSN support on Jailjhouse when deployed on a TSN-capable embedded board.
The activities may include direct collaboration with other academic and industrial partners and may reach original scientific contributions.
The worker will join a local team of developers, interns and young researchers working on similar topics or systems. |
SPHERE, other confidential parties |
Realt-Time systems, embedded systems, operating systems, networking |
Bachelor/Master |
|
|
Real-time for Xen hypervisor |
Xen is a well-known open-source hypervisor sponsored by the Linux foundation that suits not only large-scale server but also embedded and safety-critical applications. Having a mature codebase and rich feature set, Xen has been successfully chosen for several technology transfer initiatives aiming at improving its real-time capabilities.
The goal of this project is to further improve Xen real-time features, for instance by: implementing research-level memory management technologies, such as co-scheduling techniques; or improving para-virtualised driver.
The activities may include direct collaboration with Xilinx and other industrial partners and may reach original scientific contributions.
The worker will join a local team of developers, interns and young researchers working on similar topics or systems. |
Xilinx, other confidential parties |
Real-Time systems, embedded systems, operating systems |
Bachelor/Master |
|
|
Sub-meter precise positioning for autonomous driving systems |
The candidate will implement a technique for precise positioning of a robot, such as an autonomous vehicle or a drone, using one, or a mix of, the following sensors: GPS (w/o RTK correction), cameras, LiDAR, inertial+accelerometer |
|
Autonomous driving, embedded |
Bachelor/Master |
|
|
Real Time scheduling in GPU accelerated heterogeneous platforms. |
Real-time systems are those where the correctness depend on functional and non-functional timing related constraints. Programming those systems require a tight knowledge about hardware architecture, the operating systems and especially real-time scheduler, and synchronization mechanisms.
In this internship, we propose to a compiler-to-compiler tool, that allows to write real-time task code in an abstract way. Further, the real-time optimizations are achieved to generate finally a correct by construction real-time task systel in C-code, compliant to a subset of norms.
The internship student must have knowledge about object-oriented programming in C++. |
Capodieci, Zahaf |
Confidential parties |
|
Bachelor/Master |
|
Configurator system for automatic doors |
The candidate will implement a distributed system to support the design and sales department of automatic door provider |
C#, NodeJs |
Bachelor/Master |
||
|
SVR for Neural Architecture Search |
The thesis focuses on the construction of an over-parametrized Artificial Neural Network, that includes any architecture in the search space of its hyperparameters. The idea is to adopt advanced Support Vector machine for Regression (SVR) methodology to predict the ANN accuracy, to let the function to find the optimal ANN hyper-parameters for given problem or a given constraint of latency, footprint, or energy-efficiency. |
Confidential parties |
Machine Learning; Artificial Neural network, Support Vector Machines |
Bachelor/Master |
|
|
Deep Edge TinyML AutoML |
(Disponibili Intership/Progetti/Tesi di Ricerca) TinyML is to some extent about how to best implement machine learning (ML) in ultra-low power systems. Student will work on: i) Deep Learning and CNN Models Optimization for TinyML devices; ii) Deep Learning and CNN Runtime optmimization for TinyML iii) Network Architecture Search and Quantization of DL / CNN Refs: [1] https://arxiv.org/abs/2008.05124 [2] https://arxiv.org/abs/1905.13082 [3] https://ieeexplore.ieee.org/document/9049084 |
Internal Research |
IA; AutoML; NAS; TinyML; CNN; |
Bachelor/Maste |
|
|
Autonomous Nano Drones |
(Disponibili Intership/Progetti/Tesi di Ricerca) HW/SW Co-Desing of Autonomous Nano-Drones (FPGA and GPGPU). Student will work on: i) Medium/Nano Drone HW/SW Integration ii) Autonomous Navigation Algoritms Optimization (FPGA, TinyML devices) Ref: [1] https://www.bitcraze.io/2019/10/sneak-peak-of-the-ai-deck/ [2] https://www.bitcraze.io/2019/05/pulp-dronet-open-source-and-open-hardware-artificial-intelligence-for-fully-autonomous-navigation-on-crazyflie/ |
Drones; Autonomous Navigation; FPGA; TinyML |
Bachelor/Master |
||
|
RISC-V Open-Source Overlay for FPGA |
(Disponibili Intership/Progetti/Tesi di Ricerca) RISC-V based HW/SW Co-design of FPGA overlay for HW Accelerators. Student will work on: i) RISC-V Compilers; ii) HW Architecture Desing; iii) Runtime Optimization Ref: [1] https://arxiv.org/abs/1712.06497 [2] https://dl.acm.org/doi/abs/10.1145/3078659.3079071 [3] https://pulp-platform.org/ |
RISC-V; Computer Architecture; FPGA; LLVM; GCC; OpenMP; |
Bachelor/Master |
||
|
Drone perception control system |
Multiple internships are available to design HW/SW components for full- and semi-autonomous connected aerial drones. The system(s) will run on embedded platforms, such as RaspberryPi 4, and might employ machine learning techniques, possibly using the Python language and related frameworks |
Drones; Computer Vision: autonomous flight |
Bachelor/Master |
||
|
Anomaly Detection process via Machine Learning |
The thesis focuses on Anomaly Detection process in Tetra Pak data. Anomaly Detection is the process of identifying unexpected items or events in datasets, which differ from the usual behaviour. The idea is to build a machine/deep learning model using new datasets coming from the union of a machine Monitoring System and packages Visual Inspector log files. The ideal candidate is interested in Machine Learning and Artificial Intelligence and has some knowledge of the main data analysis and machine/deep learning tools (Pandas, SkLearn, PyTorch). |
|
|
||
|
Design of a planner for racing autonomous vehicles capable of overtaking at high speed |
The candidate will implement and validate, first in a simulator, then in a scale vehicle, a path planner and trajectory controller for performing overtakes and collision avoidance at high speeds for racing autonomous vehicles. If possible, it will also implement it on a real vehicle |
|
|||
|
Implementation of modern publish/subscribe middleware for Avs |
The candidate will port the driving stack for autonomous vehicles on frameworks such as ROS2/DDS, 0MQ, and explore the possible applicability of MQTT |
|
|
|
|
|
Software implementation of graph scheduling for automotive systems |
The candidate will have to tackle algorithmic problems such as allocation strategies and schedulability tests for graph tasks that model typical automotive applications. Master's (or doctoral) research thesis. |
|
Algorithms, Real-Time systems, Embedded Boards, Safety-Critical Applications |
Bachelor/Master |
FAQ
How long does it typically take to complete an internship/thesis @HiPeRT?
Since most of the projects are research projects, students will need a training period before starting the “real” work. This is why it can range from 4-6 months up to a year, depending on the project.
I want to apply for an internship (“tirocinio”) + thesis at HiPeRT. How can I do?
First of all, get in contact with the reference teacher/professor, to see if there are positions/projects available that might interest you. Then follow the instructions here, for the bureaucracy and paperwork.
Where can I find the documentation and paperwork for an internship?
For every question or doubts, and for starting the application, students might refer to the offical UNIMORE procedures (links in Italian)
- For Department of Physics, Informatics and Mathematics (FIM) –> here
- For Department of Engineering (DIEF) –> here
I am missing 4 or more exams. Should I apply for a thesis?
We typically recommend students to start the thesis when he/she is missing no more than 3-4 exams. The reason is that too many “gaps” in the period might delay the overall work!
Should I come every day in lab? Or can I work on my thesis at home?
This really depends on the technology the candidate will work on. Some thesis can be done on a “normal” laptop, other thesis, such as the ones on the F1/10 vehicle, require to have specialized hardware (the car itself!) that should not leave the lab
After graduating, I want to search for a job instead of, e.g., applying for a master course. Can you help me?
We have many contacts with companies, and we can of course support the candidate by offering projects that will be done directly in a company. For reasons of space, we did not show them, but feel free to ask!
Do you recommend to go on with master course or PhD course?
Depending on candidate’s attitude, we definitely recommend so, if one wants to push his/her skills to the limits, and learn new exciting stuff!