cv

Full Name	Seyedhamidreza Mousavi
Location	Västerås, Sweden
Languages	English (Professional), Persian (Native)
Email	seyedhamidreza.mousavi@mdu.se
GitHub	hamidmousavi0
Website	hamidmousavi0.github.io

Applied Machine Learning Researcher and Engineer with over 5 years of experience in efficient and robust deep learning. Specialized in model compression, Neural Architecture Search (NAS), robustness, and fault-tolerant learning, with strong hands-on expertise in PyTorch, distributed training, and MLOps. Proven track record of translating research into real-world deployments on GPUs, FPGAs, and microcontrollers for safety-critical and resource-constrained systems.

• 2022 – Present

  PhD in Computer Science

  Mälardalen University, Västerås, Sweden

  • Research focus on efficient, compact, robust, and reliable deep neural networks.
  • Emphasis on automatic model design (NAS) for safety-critical systems and autonomous driving.
  • Thesis:Efficient Design and Training of Compact and Robust Deep Neural Networks

• 2012 – 2019

  MSc & BSc in Computer Engineering

  Shahid Bahonar University, Kerman, Iran

  • Master’s thesis on reliability and security analysis of deep learning models.
  • Bachelor’s thesis on ARM7-TDMI implementation on Xilinx FPGA.

• 2022 – Present

  PhD Researcher – Efficient & Robust Deep Learning

  Mälardalen University, Sweden

  • Lead research on efficient and reliable deep learning for edge and embedded systems.
  • Developed one-shot NAS frameworks with knowledge distillation, achieving up to 60× reduction in search cost.
  • Designed fault-tolerant activation functions improving resilience under hardware bit-flip errors.
  • Deployed TinyDL models on STM32 microcontrollers using TinyEngine.
  • Mentored junior researchers and contributed to teaching in Deep Learning and Embedded Systems courses.

• 2020 – 2022

  Machine Learning Researcher (Remote)

  Simon Fraser University, Canada

  • Contributed to ICCAD 2020 work on aging-aware delay modeling using neural networks.
  • Conducted research on fault injection and adversarial robustness of hardware-protected DNNs.
  • Co-authored publications on stealthy bit-flip attacks and reliability-aware ML models.

• 2024

  RReLU

  • Reliable ReLU Toolbox for improving the resilience of deep neural networks under hardware faults.

• 2025

  ProARD

  • Progressive Adversarial Robustness Distillation framework for training compact and robust models.

• Machine Learning & AI: Efficient AI, Model Compression, NAS, Pruning, Quantization, Robustness, Fault Tolerance
• Frameworks: PyTorch, HuggingFace, LangGraph, TensorRT
• Distributed & MLOps: DDP, Horovod, Docker, DVC, Git, Google Vertex AI, GCP
• Programming: Python, C/C++, VHDL, LaTeX
• Embedded & Hardware: FPGA (Vivado), STM32, NVIDIA Jetson
• Soft Skills: Problem-solving, simplifying complex systems, cross-disciplinary collaboration, mentoring

• ProARD:Progressive Adversarial Robustness Distillation, IJCNN 2025
• DASS:Differentiable Architecture Search for Sparse Neural Networks, TECS 2023
• TAS:Ternarized NAS for Resource-Constrained Edge Devices, DATE 2022
• Aadam:Aging-Aware Cell Library Delay Modeling, ICCAD 2020