I am a PhD candidate at the University of Basel, Switzerland, working at the intersection of Machine Learning and Security/Privacy.

My research focuses on privacy-preserving machine learning and its applications across medical imaging and recommender systems. In particular:

• Privacy-preserving Machine Learning, especially:
  • Differential Privacy for medical imaging: Developing DP-Morph, a framework to improve the privacy–utility–performance trade-off in OCT segmentation (https://github.com/ShivaP69/DP_Morph) or (https://gitlab.com/dmi-pet-public/parsarad2025dp-morph)
  • Attacks and privacy concerns in recommender systems (gitlab.com/dmi-pet-public/Parsarad2025privacy)
  • Privacy of medical foundation models: Studying vulnerabilities and defenses in foundation models for imaging tasks.
  • Computational trade-offs in DP: Evaluating efficiency and accuracy across DPSGD variants. 
• Evaluating computational trade-offs in Differential Privacy (e.g., DPSGD variants)

Ultimately, my goal is to design methods that balance strong privacy guarantees with practical utility in real-world ML applications.

I supervise Master's and Bachelor's students on topics related to ML, security, and privacy.