Research

Topics

My research sits at the intersection of AI systems, security, and privacy. I work on making AI systems trustworthy — private, robust, auditable, and safely deployable in real-world edge and cloud environments.

Research Interests

Trustworthy AI Systems, Privacy-Preserving Machine Learning, LLM Robustness, AI Agents, Trusted Execution Environments, Confidential Computing, Anomaly Detection, Edge Computing

Postdoctoral Research

Trustworthy AI Systems

January ‘25 – Present

As a Postdoctoral Researcher at Imperial College London, I work on a coherent line of research around trustworthy AI:

  • Model privacy: membership-inference risk mitigation for deployed ML models
  • LLM robustness: prompt-leakage and jailbreak resilience
  • Agent execution: confidential/protected execution for LLM agents on edge devices
  • Agent memory: privacy-aware memory and enforced erasure for agentic AI
  • Private agent interaction: privacy-preserving protocols for agent negotiation
  • Verifiable execution: compound attestation and trusted infrastructure for AI workloads

Ph.D. Project

Efficient Anomaly Detection and Optimization in Edge Clouds

January ‘20 – February ‘24

My Ph.D. research at Umeå University focused on anomaly detection and optimization for edge computing environments. I developed methods to enhance resource efficiency and address edge clouds challenges for anomaly detection using advanced machine learning techniques, including transfer leanring, knowledge distillation, and deep neural networks.

M.Sc. Project

Improving Credit Card Fraud Detection Using Deep Learning

September ‘17 – September ‘19

The research focused on developing novel deep neural network models to address challenges of credit card fraud detection, mostly based on deep sequential models.

B.Sc. Project

Analysis of Clustering Algorithms for Streaming Data

September ‘11 – September ‘16

The research evaluated clustering algorithms for handling high-dimensional streaming data, emphasizing scalability and adaptability to real-time data processing scenarios.

Honors & Awards

YearAwardInstitution
2025EPSRC Open Plus Fellowship — Securing the Next Billion Consumer Devices on the EdgeImperial College London
2023WASP International Research Visit Grant — 150,000 SEKImperial College London
2021WASP Cluster Leader — Cloud Computing and Distributed SystemsWallenberg AI Program
2019Department Rank #1 — Computer Engineering and IT, GPA 19.14/20Amirkabir University
2019Direct Ph.D. Nomination — nominated directly without M.Sc. defenceAmirkabir University
2018Erasmus Scholarship — University Paris-Est Créteil (UPEC), FranceErasmus Programme
2017Top 0.3% — M.Sc. national entrance examIran
2011Top 0.1% — B.Sc. national entrance examIran