🌐 Links

Portfolio: https://kouchaki.com
LinkedIn: https://www.linkedin.com/in/mohammadreza-kouchaki
Google Scholar: https://scholar.google.com

Hi there, I'm Mohammadreza Kouchaki 👋

AI/ML Engineer and Computer Science graduate building machine learning systems for network intelligence, anomaly detection, and automated resource optimization in 5G/O-RAN environments. My work focuses on integrating large language models, reinforcement learning, and real-time data pipelines to create intelligent control systems for next-generation communication networks.

I design and implement AI-driven agents, ML pipelines, and anomaly detection models that operate on high-volume network telemetry and real-time infrastructure data. My work combines LLM-based reasoning, reinforcement learning, and time-series deep learning architectures to enable adaptive network optimization and automated infrastructure management.

---

🚀 Technical Focus

• AI-driven network intelligence for 5G/O-RAN environments
• LLM agents and RAG pipelines for operational decision support
• Reinforcement learning systems PPO and A2C models for dynamic resource allocation
• Real-time anomaly detection for large-scale network telemetry
• Time-series deep learning using Transformers, RNN autoencoders, GANs, and VAEs
• ML pipelines for high-volume telemetry processing
• AI-enabled orchestration in Kubernetes-based network systems

---

🧠 Current Work

AI/ML Systems – Mississippi State University

Current projects include:

• Building LLM-based meta-controller agents coordinating reinforcement learning agents for anomaly-aware network scheduling
• Developing real-time anomaly detection systems for network telemetry using Transformers and RNN autoencoders
• Implementing reinforcement learning (PPO, A2C) for dynamic radio resource allocation in O-RAN environments
• Developing RAG pipelines using LangChain/LangGraph, vector databases, and LLM reasoning
• Building O-RAN performance analyzer testbeds to evaluate AI-driven xApps and network behavior

---

🧩 Machine Learning & Engineering Stack

Programming: _{Python • MATLAB}

Machine Learning: _{Transformers • Reinforcement Learning • GANs • VAEs • RNN Autoencoders • GNNs}

ML Frameworks: _{PyTorch • TensorFlow • Keras • Hugging Face}

LLM Systems: _{RAG • LangChain • LangGraph • vLLM • Vector Databases (Chroma, Qdrant)}

Systems & Infrastructure: _{Linux • Docker • Kubernetes • CI/CD • API Development}

Data Processing: _{NumPy • Pandas • SciPy • SQL (PostgreSQL, MySQL)}

Monitoring: _{Prometheus • Grafana}

Networking: _{5G • O-RAN • IP/MPLS • SDR • Wireshark}

---

🎓 Certifications

• AWS Certified Machine Learning – Specialty
• AWS Certified Solutions Architect – Associate
• Certified Kubernetes Application Developer (CKAD)
• OCI Generative AI Certified Professional
• OCI Certified Data Science Professional
• Cisco Certified Network Associate (CCNA)

---

💼 Experiences & Projects

• Built an LLM-based meta-controller agent coordinating reinforcement learning agents for anomaly-aware network scheduling and adaptive resource allocation in O-RAN systems.
• Developed real-time anomaly detection models for network telemetry using Transformers and RNN autoencoders, improving robustness through GAN/VAE synthetic time-series augmentation.
• Implemented reinforcement learning (PPO, A2C) for dynamic radio resource allocation in Kubernetes-based O-RAN environments.
• Built RAG pipelines and LLM agent workflows using LangChain/LangGraph, Hugging Face, vLLM, and vector databases (Chroma, Qdrant).
• Designed an O-RAN performance analyzer testbed to evaluate AI-driven xApps and network behavior using large-scale telemetry data pipelines.
• Implemented multi-GPU training workflows with PyTorch and CUDA, accelerating model training for deep learning and LLM fine-tuning tasks.
• Developed lightweight AI assistants and data-analysis chatbots using LLaMA models and Hugging Face for natural language interaction with structured datasets.

• Led large-scale network optimization projects, reducing monthly operational costs by 10-15%.
• Developed performance analytics dashboards in Tableau, improving decision-making processes by 20%.
• Managed the end-to-end deployment of BH/BB solutions for large-scale B2B networks, improving network efficiency by 15%. -->

📫 Contact Me

• Email: mrkouchakii@gmail.com
• LinkedIn: linkedin.com/in/mohammadreza-kouchaki