I build things that talk to each other.
Mostly wireless networks, sensors and communication systems — because apparently regular conversations were not technical enough.
I’m Mehrshad, a Wireless Communications Engineer with a completely reasonable and definitely not concerning obsession with the invisible things keeping the modern world alive: radio waves, packets, protocols, and now, for some reason, traffic lights. I started out in Control Engineering. It took exactly one semester for me to realize that I’d much rather chase signals through the air than wrestle with feedback loops inside block diagrams. So I switched to Telecommunications and began what has now become a long-term relationship with anything involving antennas. Since then, I somehow managed to earn the unofficial title of “most papers among students” at IUST, spent five years as a TA across 13 courses, served as Head TA for most of them, and gradually discovered that I may have misunderstood the whole concept of having normal hobbies. Along the way, I worked with MTN Irancell and the Telecommunication Company, and also taught AI at Irancell Academy—because apparently radio waves alone weren’t enough to keep my schedule unnecessarily busy. Today, I work as a Project Manager at FarmadCo, managing traffic signal systems, technical teams, maintenance operations, and wireless sensor projects for data centers. So my career has gone from controlling signals, to transmitting signals, to literally managing traffic signals. At this point, I’ve stopped trying to explain the pattern and just accepted that somehow, everything in my life eventually becomes a signal-processing problem.

Mostly wireless networks, sensors and communication systems — because apparently regular conversations were not technical enough.
My work moves between AI, reinforcement learning and Python, and I’ve spent hundreds of hours teaching the same ideas without making them sound scary.
5G, 6G, wireless sensing, cryptography, network optimization — if a system is complex, measurable and slightly stubborn, I’ll probably want to understand it.
Research is great. Research that can survive real networks, real hardware and real constraints is even better.
Wireless sensor networks, 5G systems, mobility, RF electromagnetic fields, coverage, routing, energy efficiency and network reliability.
View field ↗AI-assisted radio access, QoS-aware load balancing, mobility, reinforcement learning and adaptive network control.
View field ↗Cryptographic side-channel analysis, deep-learning-assisted key recovery, blockchain security and smart-contract randomness.
View field ↗Speech and language efficiency through neural voice conversion, memory-augmented learning, graph-aware prompt compression and black-box LLM inference.
View field ↗After eight consecutive semesters of teaching programming at IUST as a Head Teaching Assistant, I delivered 260 hours of instruction across four 65-hour classes covering AI, machine-learning fundamentals, prompt engineering, AI-assisted problem solving and introductory Python.
Completed 27 teaching assistantships across communication systems, signals & systems, cryptography, electromagnetics, electronics, programming and circuits.
View TA history ↗Farmad Co L.L.C. · Tehran
Managed Tehran’s municipal traffic signal infrastructure, coordinating technical teams, maintenance operations, procurement, inventory, project delivery, and stakeholder communication to keep complex systems and ideally traffic moving smoothly. Alongside this work, led the development and deployment of wireless sensor systems for data-center applications, combining field engineering, operations, and intelligent monitoring. In short: traffic lights by day, wireless sensors by… also day.
Irancell Academy · Remote
After eight consecutive semesters of teaching programming at IUST as a Head Teaching Assistant, sometimes for two professors simultaneously, I had already become fairly comfortable explaining why the code should work when it very clearly did not. I later delivered 260 hours of AI instruction across four intensive 65-hour classes, covering Artificial Intelligence, Machine Learning, prompt engineering, AI-assisted problem solving, and Python fundamentals. My focus was on making complex ideas approachable through interactive lessons, practical exercises, real-world examples, and hands-on experimentation, teaching students not just how to use AI, but how to actually think about what it is doing.
Irancell Labs · Tehran
Managed Huawei 3G, 4G, and 5G network configuration across four major regions of Iran at Irancell, handling daily monitoring, alarm analysis, troubleshooting, optimization, site integration, neighboring-cell configuration and technical ticket resolution while coordinating with field and technical teams to keep the network stable and available. In practice, that meant staring at alarms until they confessed, fixing configuration issues before users noticed them, and making live network changes at hours when sensible people were asleep, all while keeping millions of connections running as if nothing had happened.
Iran University of Science and Technology · MBNRG
Conducted undergraduate research at IUST across Wireless Sensor Networks, Machine Learning, Reinforcement Learning, Deep Reinforcement Learning, and 5G communications, because apparently choosing just one research area would have been too easy. My thesis focused on DRL-based QoS-aware load balancing in 5G cellular networks under user mobility and uncertain channel conditions, using Python, NS-3, and mobility modeling to teach networks how to make better decisions than some humans do. Along the way, I authored and co-authored 11 research papers during my undergraduate studies, exploring AI based network optimization, wireless communications, and intelligent networking, essentially spending my bachelor’s degree turning difficult networking problems into publishable ones.
Telecommunication Company of Iran
Supported fixed-line switching operations within TCI’s telecommunications core network, diagnosing and resolving switching, signaling, connectivity, hardware, LAN, and cabling faults through system monitoring, signal analysis, and technical investigation. The role was essentially part engineering, part detective work: follow the signal, find where it disappeared, fix what broke, and restore service before anyone had time to ask why the phone was dead. It gave me hands-on exposure to core-network infrastructure, fault isolation, and real-world telecom operations, where “have you tried turning it off and on again?” is usually not an acceptable troubleshooting strategy.
R01 · Wireless Systems, Networks & RF
IET Networks · Wiley · DOI 10.1049/ntw2.70036
R04 · Intelligent Processing and Efficient AI
Preprint · July 2026 · Reinforcement Learning · Large Language Models
R03 · Security, Cryptography & Blockchain
SSRN · Posted July 2, 2026 · Blockchain · Smart Contract Security
R03 · Security, Cryptography & Blockchain
AUT Journal of Electrical Engineering · 58(Special Issue 1), 113–138 · DOI 10.22060/eej.2026.24478.5712
R04 · Intelligent Processing and Efficient AI
Results in Engineering · Elsevier · 29, 109099 · DOI 10.1016/j.rineng.2026.109099
R01 · Wireless Systems, Networks & RF
AUT Journal of Electrical Engineering · 58(Special Issue 1), 85–104 · DOI 10.22060/eej.2025.23753.5631
R02 · Intelligent Radio Access Networks
arXiv · 5G · PPO · Deep Reinforcement Learning
R01 · Wireless Systems, Networks & RF
arXiv · Mobile WSN · MARL
R01 · Wireless Systems, Networks & RF
arXiv · Wireless Sensor Networks · MARL
R01 · Wireless Systems, Networks & RF
arXiv · Game Theory · WSN
R01 · Wireless Systems, Networks & RF
International Journal of Distributed Sensor Networks · Wiley
Radio Access and Networks / RF Engineering
Telecommunication Systems & Networks
Deep Reinforcement Learning Approach to QoS-Aware Load Balancing in 5G Cellular Networks under User Mobility and Observation Uncertainty.
Seven recommendation letters across Electrical and Computer Engineering.
LET’S CONNECT
Research idea, engineering problem, collaboration, or a diagram with far too many arrows? Send the signal. I’ll do my best to decode it.