Ph.D. Candidate · AI/ML Researcher

Robust AI for security, vision, and intelligent systems.

I am a Computer Science Ph.D. candidate at Fordham University developing robust and efficient machine-learning systems for network security, computer vision, and multimodal intelligence. My research spans lightweight and noise-tolerant DDoS detection, data-efficient learning, LLM-enhanced network analysis, foreground-centric visual recognition, and vision–language generation.

My work combines empirical evaluation with research software, from data-efficient network-defense models to end-to-end visual-recognition pipelines.

Seeking academic and industry research roles · Available May 2027

Research

Focus areas

Related publications

AI for network security

Lightweight, noise-tolerant, data-efficient, and LLM-enhanced models for flow-level DDoS detection across edge, IoT, and resource-constrained networks.

Paper

Robust and data-efficient machine learning

Learning strategies for noisy, imbalanced, or limited data using augmentation, meta-learning, compact architectures, and multi-model decision fusion.

Paper

Computer vision and multimodal AI

Foreground-centric fine-grained recognition and vision–language generation that reduce background bias, capture subtle visual differences, and produce context-aware descriptions.

Paper

Featured research system

warbler_yolo

GitHub repository
YOLO11 pipeline that segments birds from field images and classifies their species
YOLO11-Seg → foreground crops → YOLO11-Cls

warbler_yolo turns unannotated field images into an end-to-end fine-grained recognition workflow. It segments and crops foreground subjects, creates standardized data splits, then trains and evaluates a YOLO11 classifier through a command-line pipeline.

Prepare · Segment · Crop · Split · Train · Evaluate

Academic research

Selected peer-reviewed publications

ShallowNet: A Lightweight Neural Network Approach for Efficient Flow-Level DDoS Detection

Ali Alfatemi, Mohamed Rahouti, Abdellah Chehri, Zakirul Alam Bhuiyan

IEEE Transactions on Network and Service Management · Vol. 23, pp. 5940–5949

PaperDOI

A Two-Stage LLM-Enhanced DDoS Detection Framework for Next-Generation IoT and Edge Networks

Ali Alfatemi, Mohamed Rahouti, Zakirul Alam Bhuiyan, Abdellah Chehri, Aiman Solyman

IEEE GLOBECOM 2025 · CISS Symposium, pp. 7–12

PaperDOI

Vision-Language Integration for Image Captioning Using Vision Transformers and GPT-J

Ali Alfatemi, Mohamed Rahouti, Mohammed Aledhari, Nasir Ghani, Abdellah Chehri, Gwanggil Jeon

Applied Imagery Pattern Recognition · LNCS 16446, pp. 101–114

PaperDOI

Foreground-Centric learning improves robustness in fine-grained visual recognition

Ali Alfatemi, Mohamed Rahouti, Majjed Al-Qatf, Senthil Kumar Jagatheesaperumal

Signal, Image and Video Processing

PaperDOICode

Academic record

Education, appointments, and service

Ph.D.
Computer ScienceFordham University · Expected May 2027
M.S.
Computer ScienceSouth China University of Technology · 2021
Appointment
Teaching FellowFordham University · 2025–present
Teaching
CISC 1100: Structures of Computer ScienceFordham University · Fall 2025 and Spring 2026 · Syllabi: Fall 2025 PDF · Spring 2026 PDF
Fellowship
AI Summer Research FellowFordham University · Summer 2025
Internship
Student Research InternTsinghua Shenzhen International Graduate School · 2021–2022
Service
Peer reviewerIEEE TNSM · Information Fusion · Scientific Reports · IJCNN 2025
Patent
CN113537358B · InventorCancer subtype identification via multi-omics data integration · Patent record ↗

Contact

Research collaboration and opportunities

I welcome focused conversations about academic opportunities, research collaboration, and applied AI roles.