During 2013-2020 I have been a PhD Student, at KTH - Royal Institute of Technology in Stockholm, Sweden. I worked in the Division of Electronics and Embedded Systems at the School of Electrical Engineering and Computer Science.
I worked under the supervision of Ingo Sander (2013-2020), Christian Schulte (2016-2020), and Werner Zapka (2013-2015).
About my research
ForSyDe (Formal System Design) is a methodology providing modeling and refinement techniques that can be used as part of the design process of SoCs. Founded by Ingo_Sander, it has become an open-source initiative.
ForSyDe-Atom is a spin-off of the ForSyDe-Shallow DSL, with focus on: 1) orthogonalization of concerns; 2) deconstruction of semantics to their core; 3) primitive interaction.
DeSyDe is a design space exploration tool for system-level optimization of heterogeneous platforms shared by multiple applications, founded by Kathrin Rosvall.
CoInSyDe is a multi-target component-based template expander used for synthesizing compilable code from solved (e.g. mapped, scheduled) application models.
f2cc (ForSyDe-to-CUDA C) is a software synthesis tool developed as part of the ForSyDe design flow, carrying on the work of Gabriel Hjort Blindell
In partnership with Xaar I was involved in the development of an on-line reliability system for industrial printheads, having high throughput demand and real-time constraints.
Most resources spent in designing a real-time, especially safety-critical system, are spent during validation. This is why in our research group we opt for a "correct-by-construction" approach on system design
Putting different stuff together in the same system is fun, but there are chances that things blow up. Therefore we need to understand systems both as a whole and as individual units trying to understand each other.
Combining continuous dynamics and different computing models almost always ends up in chaotic behavior. This is because different parts of a system lack a common understanding about what time actually means.
The abrupt shift towards parallel computing has left many gaps in understanding how thing actually work. This is where us, researchers, come into the picture! [check Ch.3 of my MSc Thesis for more...]
This paradigm lets you describe programs as abstractions rather than as instructions. Combined with a powerful algebra of types, this is an awesome tool for system design.
It is a growing endeavour to make machines understand what we want them to do. In the context of EDA, I am tackling this by trying to find the minimal set of language constructs for describing and synthesizing systems.
CASTOR: Correct-by-Construction Seed project leading to a pre-study aiming to identify how formal verification techniques can be further integrated into ForSyDe.
CORRECT Swedish national project which aims to create a full correct-by-construction design flow, including design space exploration, down to software generation.
SAFEPOWER EU project whose objective is to enable the development of cross-domain mixed-criticality systems with low power, safety and security requirements.
CONTREX EU project aiming to enable energy efficient design through analysis and optimisation with regard to application demands at different criticality levels.
iPack VINN Excellence Center aspiring to advance the Swedish market for intelligent packaging by developing innovative electronics in vision of internet-of-things.
|
|
|
|
MSc
|
|
|
BSc
|
|
