News
-
Our paper EFU: Enforcing Federated Unlearning via Functional Encryption is now available in the ACM CIKM 2025 proceedings.
-
Presented EFU: Enforcing Federated Unlearning via Functional Encryption at CIKM 2025 in Seoul, Korea.
-
Our paper Empirical Analysis of Asynchronous Federated Learning on Heterogeneous Devices: Efficiency, Fairness, and Privacy Trade-offs is now available in the IEEE IJCNN 2025 proceedings.
Research & Publications
Privacy-Preserving & Communication-Efficient FL
-
CIKM 2025 — 34th ACM Conference on Information and Knowledge Management.
-
PMCJ 2025 — Pervasive and Mobile Computing Journal.
-
NeurIPS 2024 — Workshop on Federated Foundation Models.
-
ISPEC 2023 — Conference on Information Security Practice and Experience.
Efficiency, Fairness & Privacy Trade-offs
-
IJCNN 2025 — International Joint Conference on Neural Networks.
-
FedCSIS 2023 — 18th Conference on Computer Science and Intelligent Systems. Best Paper Award.
Survey
-
JPDC 2024 — Journal of Parallel and Distributed Computing.
Professional Experience
-
- Designed a group-aware architecture for fair and private cross-silo federated learning.
- Developing efficient and fair LLM fine-tuning in federated learning using Low-Rank Adaptation (LoRA) — ongoing.
-
Research across two EU projects — DAIS (a 47-partner EU flagship) and DADAP — with academic, industrial, and clinical partners.
- Built a physical federated learning testbed for Speech Emotion Recognition across heterogeneous edge devices, benchmarking efficiency, fairness, and privacy across hardware tiers.
- Developed a scalable, efficient encryption method for FL that compresses model updates — cutting communication and computation cost while preserving privacy.
- Developed and benchmarked modern tabular architectures on clinical mental-health data, improving predictive performance and fairness for deployment in Swedish hospitals.
-
- Software development — developed new modules on the Odoo ERP platform, integrated with existing finance and supply-chain modules.
-
- Developed a graph inductive-learning approach for anomaly detection on dynamic information networks, improving detection accuracy on evolving graph structures.
Education
-
University of EdinburghPh.D. in Cyber Security, Privacy, and Trust.
Thesis: Efficient and Trustworthy Federated Learning for Privacy-Critical Applications. -
University of TehranM.Sc. in Information Technology Engineering.
Thesis: Anomaly Detection in Dynamic Information Networks. Grade: A.
Talks & Teaching
Talks & Presentations
-
Paper TalkCIKM 2025 — 34th ACM Conference on Information and Knowledge Management. Seoul, South Korea · Nov 2025.
-
Paper TalkIJCNN 2025 — International Joint Conference on Neural Networks. Rome, Italy · Jun 2025.
-
Paper TalkNeurIPS 2024 — Conference on Neural Information Processing Systems. Vancouver, Canada · Dec 2024.
-
Invited TalkFlower Monthly — “EncCluster: Scalable Functional Encryption in Federated Learning,” hosted by Prof. Nicholas Lane (Univ. of Cambridge / Flower Labs) · Sep 2024. Recording.
-
Invited TalkUniversity of Tartu — 2nd Summer School on MegaData: Federated Machine Learning. “Balancing Privacy and Performance in Federated Learning” · Aug 2024.
-
Paper TalkFedCSIS 2023 — 18th Conference on Computer Science and Intelligent Systems. Warsaw, Poland · Sep 2023.
-
Paper TalkISPEC 2023 — 18th Conference on Information Security Practice and Experience. Copenhagen, Denmark · Aug 2023.
Teaching
-
Teaching AssistantComputational Data Mining — University of Tehran · 2020.
-
Teaching AssistantComplex Networks — University of Tehran · 2019–2020.
Program Committee
-
CFAgentic @ ICML’25 — ICML 2025 Workshop on Collaborative and Federated Agentic Workflows.