Mohammadreza Kouchaki AI / ML Engineer for Network Intelligence

Building AI systems for network intelligence and cloud-native environments, reinforcement learning optimization, anomaly detection, and LLM agents.

Certifications

Click any credential to view the official badge.

AWS Certified Solutions Architect – Associate

Amazon Web Services (AWS) · 2024

Credential

AWS Certified Machine Learning – Specialty

Amazon Web Services (AWS) · 2024

Credential

OCI Generative AI Certified Professional

Oracle University · 2025

Credential

OCI Certified Data Science Professional

Oracle University · 2025

Credential

Cisco Certified Network Associate (CCNA)

Cisco · 2025

Credential

Certified Kubernetes Application Developer (CKAD)

CNCF / Linux Foundation · 2025

Credential

Projects

Selected Projects

Reproducible ML systems, LLM agents, and network intelligence work.

RAG Research Agent (Chroma + LangGraph + Tools)

KB-first RAG agent built with LangChain and LangGraph, featuring vector retrieval, external search tool routings, lightweight UI, and structured API responses.

LLM Agents RAG Tool Routing

Stack: Python · FastAPI · ChromaDB · SentenceTransformers · vLLM

GitHub Live Demo

Actor-Critic xApp for Reinforcement Learning-based Resource Optimization

Reinforcement learning policy for resource scheduling; metrics tracked via Prometheus/Grafana.

Reinforcement Learning Optimization O-RAN

Stack: PyTorch · RL · Kubernetes · Prometheus · Grafana

Project Page GitHub Paper / Write-up

RAG Pipeline with Qdrant for Medical Knowledge

Domain-specific RAG pipeline using Qdrant for semantic retrieval over medical knowledge, paired with a local LLM for grounded question answering.

RAG Vector Search Qdrant LLM Systems

Stack: Python · Qdrant · SentenceTransformers · Transformers · Pandas

GitHub

Time-Series Anomaly Detection

RNN-based Variational Autoencoder and PyTorch Designed Transformer for Real-time time-series anomaly detection for network signals with attention-augmented sequence models.

Anomaly Detection Time-Series PyTorch 5G

Stack: PyTorch · Transformers · Kafka · Docker

Project Page GitHub Paper / Write-up

Adaptive Cross-Attention Multimodal Graph-LLM for Social Network Threat Detection

Research prototype integrating PyTorch based Graph Transformer and Encoder-Based LLM embeddings through adaptive cross-attention to jointly reason over user interaction graphs and tweet text for bot detection.

GraphLLM Multimodal Learning Bot Detection Social Network Analysis

Stack: PyTorch · DGL · Transformers · LLM-Encoding · Graph Neural Networks

GitHub

Live Demos

Interactive demos hosted on Hugging Face Spaces.

RAG Research Agent Demo

Hugging Face Spaces

Ask questions and see KB-first retrieval with web fallback.

RAG LLM Agents

Launch Demo

Anomaly Detection Visualization

Hugging Face Spaces

Visualize model outputs and detection patterns on time-series signals.

Time-Series Anomaly Detection

Launch Demo

HF profile: huggingface.co/mrkouchaki

Publications

Selected papers and write-ups.

Federated Neuroevolution O-RAN: Enhancing the Robustness of Deep Reinforcement Learning xApps
IEEE Communications Magazine · 2025
LLM-Based Meta-Controller for Anomaly-Aware O-RAN Network Slicing
IEEE Transactions on Wireless Communications · 2026
O-RAN Performance Analyzer: Platform Design, Development, and Deployment
IEEE Communications Magazine · 2024
Actor-Critic Network for O-RAN Resource Allocation: xApp Design, Deployment, and Analysis
IEEE GLOBECOM · 2022
View paper
Advancing O-RAN Security: Integrated Intrusion Detection and Secure Slicing xApps
IEEE INFOCOM Workshops · 2024
LLM-Guided RL: A Reasoning-Augmented Framework for Intelligent O-RAN Network Slicing
IEEE ICC · 2026
Enhanced Real-Time Threat Detection in 5G Networks: A Self-Attention RNN Autoencoder Approach for Intrusion Analysis
IEEE 22nd WiOpt · 2024
Adaptive Cross-Attention-Based Multimodal GraphLLM for Social Network Threat Detection
arXiv e-prints · 2025
OpenAI dApp: An Open AI Platform for Distributed Federated Reinforcement Learning Apps in O-RAN
IEEE Future Networks World Forum (FNWF) · 2023
Experiment on High Capacity Backhaul Transmission Link Aggregation Solution for 5G Networks
IEEE Conference of Open Innovations Association (FRUCT) · 2020
Retinal Identification System Based on the Combination of Fourier and Wavelet Transform
Journal of Signal and Information Processing · 2012
Down-Conversion Self-Oscillating Mixer by Using CMOS Technology
IEEE International Colloquium on Signal Processing · 2012
Design and Simulation of Low-Noise Bandpass Filter by Using Active Capacitance Circuit
IEEE International Colloquium on Signal Processing · 2012
A 6-bit Digital Phase Shifter by Using HEMT Technology
IEEE International Colloquium on Signal Processing · 2012

Google Scholar

Technical Skills

Core technologies and systems I use to build AI, machine learning, and network intelligence solutions.

Machine Learning & Deep Learning

Supervised & Unsupervised Learning, Reinforcement Learning (DRL), Large Language Models (LLM), Autoencoders, Transformers, CNN, LSTM, Graph Neural Networks (GNN).

Programming

Python (TensorFlow, PyTorch, Keras), MATLAB for numerical computing, deep learning model development, and scientific computing.

Systems & Infrastructure

Linux environments, Kubernetes orchestration, Docker containers, CI/CD pipelines (GitHub Actions), CUDA acceleration, and Software Defined Radio (SDR).

Data Processing

NumPy, SciPy, Pandas for large-scale data processing, feature engineering, and analytical pipelines.

Cloud Platforms

AWS, Google Cloud Platform (GCP), Microsoft Azure, and Oracle Cloud for AI deployment and scalable infrastructure.

Networking & Telecom

4G, 5G, and O-RAN architectures, RAN/backhaul systems, Huawei RTN, OTN, Ericsson MINI-LINK, and BTS 3900.