Research

My research interests can be generalized to these four fields:

  • Heterogeneity (Accelerators for Machine Learning and Cryptography, SoCs, GPGPUs, Task scheduling)
  • (Domain-Specific) Programming languages (expressing and automatically detecting parallelism, automatic detection of the most suitable platform type (CPU, GPGPU, FPGA) for code execution)
  • Parallelism (processing, programming, programming models)
  • Functional / Reactive programming (Programming highly heterogeneous and parallel hardware, Actor model, elasticity, resiliency, generic software engineering using functional programming languages)

Research projects:

  • HCSCA - Heterogeneous Computing Systems with Customized Accelerators (July 2019 - July 2022)
    • Role: Researcher
    • ETH Zurich and UniZG FER joint project which aims to develop a novel heterogeneous accelerated computing system architecture based on RISC-V with accelerators for disruptive application domains (namely Machine learning, Crypto processing, Video processing). Furthermore, the project includes software support for proposed heterogeneous architecture and validation of disruptive applications. The project is funded by Swiss-Croatian Cooperation programme (Swiss National Science Foundation (http://www.snf.ch/) and Croatian Science Foundation (https://www.hrzz.hr)).
  • Hazelnut
    • Role: Researcher
    • This project is a joint project between distinct researchers at UniZG FER and Green Light Technologies Ltd. Project aims to develop a low cost ultra energy efficient modules for wide variety of Internet of Things applications by exploiting heterogeneity using cheap but reliable off the shelf components.
    • Paper
  • Jobbee
    • Role: Researcher / Developer
    • An intelligent matchmaker between generic job offerers and clients.
  • Cool Cracker Cluster
    • Role: Researcher
    • This research project by UniZG FER was investigating on how to exploit heterogeneity to raise energy efficiency and performance in bcrypt hash computation. The ulimate deliverable of this project is our Cool Cracker Cluster.
    • Paper
  • Houseleek
    • Role: Consultant
    • Houseleek is a project which aims to deliver a novel reliable and centralized facility management system. By utilizing Deep Neural Network to authorize entrance to individual spaces using image of the user’s face, project Houseleek introduces new ergonomy standards in large and managed premises.
    • Paper

Research activities:

  • ACACES 2019 - Fifteenth International Summer School on Advanced Computer Architecture and Compilation for High-Performance and Embedded Systems, July 2019, Fiuggi, Italy
  • Academic visitation - professor Philip A. Wilsey, March 2019, Cincinnati, USA
  • Academic visitation - professor Lennart Johnsson’s research group, February 2018, Houston, USA

Grants:

  • USENIX Student Grant for USENIX Security Symposium (12. 8. 2019. - 16. 8. 2019., Santa Clara, CA, USA)
  • HiPEAC Collaboration Grant (July 2019.)
  • Academic Mobility Grant by University of Zagreb to visit the University of Houston (21. 2. 2019. - 1. 3. 2019., Houston, TX, USA)
  • Academic Mobility Grant by University of Zagreb to visit the University of Cincinnati (11. 3. 2019 - 17. 3. 2019., Cincinnati, OH, USA)
  • HiPEAC Travel Grant to visit the HiPEAC Conference in Valencia (21. 1. 2019. - 23. 1. 2019., Valencia, Spain)
  • HiPEAC Grant for the ACACES 2019 Summer school (14. 7. 2019. - 20. 7. 2019., Fiuggi, Italy)