Major Projects
•AI for anomaly detection in smart factory data, 2020-2021
This master thesis project has been defined between Software Competence Center Hagenberg-Austria and
KTH-Sweden. from the KTH side, Professor Mihhail Matskin is the project coordinator and thesis examiner.
In this project, we aim at designing AI method and tools, that are context-aware, for consistency check of
factory data, and discovering potential anomaly in the captured data.
•AI-powered side-channel attack to hardware systems, Fall 2019.
Deep learning side-channel attacks are significant security threats for cryptographic algorithms implementa-
tions. In this project, we aim at studying performance impacts of changing the testing interval of the captured
power traces on the success rate in the key recovery attempts. The early stage results show that by section-
wise sliding of power traces no false positive results are introduced. The results further show that retrieving
a value from a specific keybyte position needs that specific model for that specific keybyte position.
•AI-powered fraud detection in credit card transactions, Fall 2019.
In this project, a Generative Adversarial Networks (GAN) model was used for the classification of credit card
fraudulent transactions from a highly imbalanced set of credit card transactions (very few frauds among a
huge amount of trustworthy transactions).
•Building networked systems security, Jan 2020.
In this project, we have designed and implemented a solution for securely extending IT work-space of a
tech company to different locations. The solution has been proposed based on the needs of company, and
includes implementation of custom firewalls, and authentication approaches. The solution has been finally
tested over real network components, i.e. routers/switches and end-users’ laptops.
•AR/AI-based instruction aider for assembling circuits, Fall 2018.
In this project, we designed and implemented an application for assisting users in assembly of electrical
circuits. Our solution leverages augmented reality (for communications with user), AI (for detection of com-
ponents), and multimodal interactions (for HCI) for bringing safety, time and cost efficiency to the system.
In our solution, the electric circuit are first identified, and then using augmented reality, user gets instruction
where to put each component. The final system was evaluated by users in practice, and from their survey,
the instruction aider has reduced the time taken about 19.65% to assemble the circuit. Furthermore, all the
users have preferred using the application than reading the instructions from the manual due to its ease of
use.
•MIniature STudent satellite: Embedded software development for synchronization, Fall 2019.
The MIST satellite is under development by a student run team at KTH. The two tasks implemented in the
project were time synchronization between the ground station and satellite, and improving telecommand
handling. The developed codes for time synchronization and telecommand handling were integrated with
the core of project, and promising results were observed from the tests.
•Real-time cruise control applications on real-time operating systems, Fall 2018.
In this project, carrying out real-time applications on typical real-time operating systems was investigated
in order to explore real-time application’s performance and memory footprint trade-offs. More specifically,
a cruise control application using the DE2/DE-115 board was implemented, and the synchronization and
communication between tasks in real-time were carried out.
•Image tracking using MicroC/OS-II, Winter 2019.
In this project, we designed an image tracking algorithm tracking a given moving ”circle” pattern in a series
of image frames for the purpose of further processing. We implemented this application on a given custom
multi-processor hardware platform hosted on a DE2 FPGA, in three ways: (i) on a single core, (ii) using the
MicroC/OS-II on a single core, (iii) and without OS on multiple cores. The aim of this project was shedding
light onto the complexity of programming over multi cores and understanding the required communications
in parallel processing for delivering a service.
Other Projects
• Sound and Gestures in Interaction: In this project, we used laptops and smartphones to map connections
between physical movement and sound. The embedded sensors on the smartphone were used to detect and
differentiate between different gestural actions and connect those actions in a meaningful way to a set of
provided sound models.
• White/black box testing, requirement and model based testing, and automated testing, Winter 2019.
• Implementation of a Swedish Event Planning software by using extreme programming, Fall 2018.
• Design and implementation of a paper review database, Spring 2011.
• Design and implementation of an archive system for FUM University, 2010.
• Programming Pac-Man Game, Spring 2011.
• Programming multi Sudoku, Spring 2011.
• Programming Missionaries and Cannibals Problem with prolog language, Spring 2011.
• Programming Galcon game, Spring 2011.
• Programming a line following robot, Spring 2009.
• Programming a multi-task dictionary, Spring 2009.