Research Assistant, KU 6G Research Center
Amine Kidane
Agentic AI-assisted orchestration for 6G vehicular and edge-cloud systems.
I build Agentic AI-assisted systems for 6G vehicular networks, edge intelligence, and real-time service orchestration. My current work sits where multimodal context, retrieval-augmented reasoning, and reinforcement learning meet the practical constraints of cloud-edge infrastructure.
Abu Dhabi, United Arab Emirates
Khalifa University
6G Research Center
Publications
Research Output
Move-LLM: Multimodal LLM-Assisted DRL Framework for On-Demand Service Deployment in 6G Vehicular Networks
IEEE Transactions on Vehicular Technology
Multimodal LLMsDRL6G vehicular networksOn-demand deployment
Move-LLM studies on-demand service deployment for 6G vehicular networks using multimodal perception and DRL-based placement decisions. The architecture connects rich context from vehicles and infrastructure to a learning-based deployment policy that can react to mobile, resource-constrained network conditions.
Main Contributions
- Uses multimodal signals to improve context awareness for service deployment decisions.
- Couples LLM-assisted semantic reasoning with DRL policy learning for dynamic vehicular environments.
- Targets low-latency, resource-aware deployment of services across 6G vehicular edge infrastructure.
AI-Native Cloud-Edge Orchestration for 6G Metaverse Networks: An LLM-Guided Multi-Agent DRL Approach
Complex & Intelligent Systems, 12(4), Article 141
AI-native orchestrationTopoRAGMulti-agent DRLMetaverse networks
This work proposes an AI-native orchestration framework for 6G metaverse services. An LLM-based cloud planner uses topology-aware RAG to generate global soft priors, while decentralized edge DRL agents adapt those plans under local resource and mobility conditions.
Main Contributions
- Introduces an LLM cloud planner that uses TopoRAG over historical deployment traces.
- Streams semantic priors to decentralized DRL agents for real-time edge orchestration.
- Uses deviation-based rewards to align local execution with LLM-predicted deployment costs.
On-Demand Vehicular Service Deployment in 6G: A Collaborative Large-Small LM Architecture
IEEE Vehicular Technology Magazine
Large-small LM collaborationVehicular servicesEdge intelligence6G
This magazine paper frames on-demand vehicular service deployment as a collaborative large-small language model problem. Large models provide global service understanding and planning, while smaller edge-side models support fast, local adaptation under vehicular mobility and resource constraints.
Main Contributions
- Defines a collaborative large-small LM architecture for vehicular service deployment.
- Separates global semantic planning from fast local execution near the network edge.
- Positions LM collaboration as a practical design path for 6G vehicular intelligence.
A RAG-Assisted DRL Framework for Microservices Deployment in 6G Vehicular Networks
IEEE WiMob 2025
RAGDRLMicroservice DAGsEdge-cloud orchestration
This paper integrates a lightweight DRL deployment agent with a graph-based RAG module powered by a partially frozen LLM. The RAG layer retrieves relevant deployment history and service intent, then turns it into soft placement guidance and reward estimates for real-time microservice deployment.
Main Contributions
- Adds semantic awareness to DRL scheduling through graph-based retrieval and LLM planning.
- Models deployment memory with resources, network latencies, SLA feedback, and service context.
- Improves convergence and generalization for microservice DAG placement under changing 6G edge conditions.
Towards On-Demand Metaverse Service Deployment in 6G Vehicular Networks Using Multimodal LLMs
IEEE INFOCOM 2025 Workshops
Multimodal LLMsMetaverse servicesOBU clustersQoS-aware selection
This work proposes a vehicular MLLM-driven framework for context-aware metaverse service deployment in 6G vehicular networks. It uses text, image, and video context to recommend services, then selects RSU or OBU-cluster hosting nodes based on QoS, resource utilization, and resource constraints.
Main Contributions
- Uses multimodal context to improve service recommendation for vehicular metaverse workloads.
- Introduces OBU clusters as flexible deployment options beyond fixed RSU locations.
- Develops QoS- and resource-aware node selection for on-demand service placement.
A Hierarchical Agentic Framework for Self-Organizing Next Generation Wireless Networks
IEEE Network Magazine
Under-review manuscript on hierarchical agentic control for self-organizing next-generation wireless networks.
Agentic networksSelf-organizationWireless intelligence
Large Language Model-Based Trajectory Forecasting for Dependency-Aware Microservice Migration in 6G Vehicular Edge Networks
IEEE Journal on Selected Areas in Communications
Under-review manuscript on LLM-based trajectory forecasting for dependency-aware microservice migration.
Trajectory forecastingMicroservice migrationVehicular edge networks
CV
Academic Profile
Education, research experience, technical skills, selected projects, awards, and service in one continuous academic view.
Education
BSc in Computer Engineering, AI Track
Aug 2019 - Jan 2024Khalifa University - Abu Dhabi, UAE
- GPA: 3.96/4.0
- President's List: 8/9 semesters
- Full Undergraduate International President's Scholarship
Experience
Research Assistant
Jun 2024 - PresentKU 6G Research Center, Khalifa University - Abu Dhabi, UAE
Research on Agentic AI-assisted orchestration, RAG/DRL microservice deployment, and multimodal intelligence for 6G vehicular edge systems.
- Built publication-track frameworks spanning multimodal LLM reasoning, RAG-enhanced DRL, and edge-cloud deployment.
- Model service placement as a real-time decision problem across vehicular mobility, QoS, resource constraints, and network state.
- Collaborate across KU, C2PS, XJTLU, LAU, Concordia, and Carleton research networks.
AI Software Engineer Intern
Jan 2024 - May 2024Rhetora.ai - Remote
Built an AI co-pilot sales platform across prospecting, lead nurturing, and deal closure workflows.
- Integrated automation, scheduling, and AI assistance into applied sales workflows.
- Focused on practical decision-support tooling rather than isolated model demos.
Visiting Undergraduate Researcher
May 2023 - Aug 2023eBrain Lab, NYU Abu Dhabi - Abu Dhabi, UAE
Designed a quantized and hardware-aware neural architecture search framework for autonomous robots.
- Achieved 15% faster inference and 1.2x higher energy efficiency on embedded GPUs.
- Explored hardware-software co-design for deployable AI in robotic settings.
Cybersecurity Intern
Jun 2022 - Sep 2022EBTIC - Abu Dhabi, UAE
Studied defenses against adversarial attacks on machine-learning models in cyber-physical systems.
- Evaluated robustness-enhancement strategies for safety-sensitive settings.
- Improved resilience with a 5.7% reduction in successful attack rates.
Technical Skills
Research
Agentic integrationRAGDRL6G networksVehicular edgeMicroservice orchestration
Languages
PythonC++CJavaJavaScriptTypeScriptMATLABSQL
ML Frameworks
PyTorchTensorFlowHuggingFaceOpenCVScikit-learnKeras
Networking
SUMONS3Omnet++VeinsINET
Systems
HPCAWSAlibaba CloudDockerROS 2PX4
Databases
PostgreSQLMySQLMongoDBLLaMAIndex
Projects, Awards, and Leadership
Selected Projects
- Automatic Waste Segregation System. Senior design project combining computer vision and mechanical actuation for smart waste sorting.
- Electrical Impedance Tomography Image Reconstruction. Variational autoencoder research for ill-posed inverse problems in medical imaging.
- MBZIRC Maritime Grand Challenge. Computer vision and navigation for heterogeneous UAV and USV teams in GNSS-denied environments.
- ZenMate. Alibaba Cloud GenAI Hackathon chatbot built with Llama 2.
- Quantum Crop Rotation Optimizer. Quantum-computing-based crop rotation optimizer for UAE agriculture.
- High-Speed Printing Anomaly Detector. Industrial anomaly detector using PaDiM for high-speed printing systems.
Awards
- EDGE Pioneers 4.0 Hackathon Series 3
- Alibaba Cloud GenAI Hackathon 2023 - Sustainability Expert in GenAI for Social Good in MENA
- 17th IEEE Software Engineering Competition - Second Prize
- Khalifa University Programming Contest - First Prize, twice
- Musabaqat Mathematics Competition V - 5th among 90 students
- Open Mathematics Olympiad - Second Prize
- Golden Key International Honor Society Inductee
- Co-founded the AI Club at Khalifa University and helped grow it to 250+ members.
- Completed 150 hours of peer tutoring in C++ and Calculus.
Contact
Contact details
ProfilesKU GitHub LinkedIn Google Scholar