Research Student
- Conducted research aimed at tackling the challenges in next-generation wireless networks.
- Implemented deep reinforcement learning techniques for spectrum management in wireless networks.
Oluwaseyi Giwa
Welcome
giwaoluwaseyi130@gmail.com
+234 (703) 586-7514
About Me
I am interested in research on next-generation wireless networks. My master's thesis was on resource allocation in heterogeneous wireless networks using deep reinforcement learning. I was fortunate to be
advised by Professor Jonathan Shock and
Dr Jaco du Toit at the University of Cape Town.
In addition to my Master's program, I was fortunate to be advised by Dr Muhammad Ali Jamshed from
the University of Glasgow, on a couple of research projects.
Prior to this, I was an undergraduate student at Olabisi Onabanjo University (OOU),
where I graduated with a Bachelor's degree in Electrical and Electronics Engineering, advised by
Dr Peter Alao
and Dr Matthew Olajide
at the Wireless Communications Lab. My final year thesis was on the design and development of a
full-duplex wireless transceivers system using 433MHz circuit modules.
The emergence of 6G network would see more integration of AI in its architecture and functionality. While the promise of 5G
network were not fully met, research on 6G networks has shown that we can achieve a major part of the promise. Thus, my research directions include:
(1) Safe and efficient reinforcement learning for wireless networks. The traditional network optimization methods often assume fixed network conditions and
static users equipment, limiting their practical applications in real-world network scenarios. While several literatures have studied machine learning techniques, particularly, reinforcement learning as a tool to replace these traditional methods,
there are still concerns regarding the safety of RL application to wireless networks due to its dynamic and non-convex nature. I aim to improve the safety and efficiency of RL application in wireless environment by
utilizing constrained Markov decision process and Bayesian neural networks to guide RL agents from high-risk states and mask unsafe action in these wireless environments.
Presently, I have proposed two RL architectures that address these concerns, Meta RL and QPPG.
(2) AI-native wireless networks. An AI-native 6G network leverages AI techniques (e.g. ML, DL, neural networks, etc) for the design,
deployment, management, and operation of various network and device functions. AI-native is important because of its adaptability, handling non-linear signals, and resource efficiency.
I aim to improve the problem of limited generalizability, limited interpretability, reduce training duration, and mitigate unsustainable computing process.
My work investigates how tranfer and continual learning could be used to address the issue of generalizability and frequent retraining. Second, integrating LLMs in network controller to
achieve improved interpretability.
Please feel free to reach out to me to share your thoughts or explore any form of collaboration!
News
-
[September 2025] Our paper on LLM-RAN got accepted at NeurIPS 2025 Workshop.
-
[August 2025] Served as a reviewer for IEEE Transactions on Circuits and Systems for Video Technology.
-
[July 2025] Graduated with a distinction in MSc from the University of Cape Town!!!
-
[July 2025] Got selected for Black in AI emerging leaders in AI program.
- Developed a Gaussian diffusion model with a transformer-based denoiser to generate human motion sequences from sparse input key poses.
- Trained the model on an extensive motion capture database of various natural human activities from the Archive of Motion Capture as Surface Shapes (AMASS).
- Designed a composite geometric loss function to minimise position and velocity losses of joints in the kinematic chain.
- Enhanced an existing C++ codebase to improve debit card readability on point-of-sale (POS) machines, resulting in seamless financial card transactions for various clients.
- Utilised Postman software to manage transactions requests from various POS machines, contributing to a revenue increase for the firm in the first two quarters of 2024.
- Developed a 25-watt wireless transceiver using the FS1000A transmitter module, capable of transmitting signals at 92.7 MHz within a 100-meter radius, effectively minimising signal interference.
- Designed a 12dB standard gain end-fire array antenna using MATLAB and integrated it into a 5G mmWave radio device, increasing the radio's directivity by 166%.
-
Quantum-Preconditioned Policy Gradient for Link Adaptation in Rayleigh Fading Channels
-
Meta-Reinforcement Learning for Fast and Data-Efficient Spectrum Allocation in Dynamic Wireless Networks
-
Design and Development of Full-Duplex Wireless Transceivers System using 433MHz Circuit ModulesM. Olajide, O. Giwa, J. Ogunsakin, O. Aina, D. Kuponiyi
FUW Trends in Science & Technology Journal [published, 2025] [PDF] -
Optimisation of Resource Allocation in Heterogeneous Wireless Networks using Deep Reinforcement Learning
-
Dynamic Causal Discovery of Temporal Dependencies in Wireless NetworksO. Giwa, C.K. Thomas
IEEE INFOCOM [under review, 2026] [Code] -
The LLM as a Network Operator: A Vision for Generative AI in the 6G Radio Access NetworkO. Giwa, M. Adewole, T. Awodumila, P. Aderinto
NeurIPS 2025 Workshop on AI and ML for Next-Generation Wireless Communication and Networking [poster, 2025] [PDF] -
Human Motion Synthesis: A Diffusion Approach for Motion Stitching and In-Betweening.M. Adewole, O. Giwa, F. Nerrise, M. Osifeko, and A. Oyedeji
[PDF] - Volunteer – African Institute for Mathematical Sciences Schools Enrichment Centre (2025).
- President – National Institute of Electrical and Electronics Engineering Students' Association (2022)
- Mentor – Ibogun Campus Tech Community (2023)
- Google DeepMind $2000 research support ( 2025)
- Mastercard Foundation Scholarship for postgraduate studies (2024 to 2025)
- International Scholarship Initiative recipient (2024)
- Department of Electrical Engineering best graduating student (2024)
- Federal Government of Nigeria Scholarship for undergraduate studies (2020 to 2023)
Research & Work Experience
Research Assistant
Backend Developer Intern
Research Assistant
Publications
Journals
Conferences
Workshops
Preprints
Service
Awards & Accolades
Education
-
MSc. in Mathematical SciencesUniversity of Cape TownFeb. 2024 - July 2025Grade: Distinction
-
B.E. in Electrical and Electronics EngineeringOlabisi Onabanjo University (OOU)Oct. 2017 - May 2023CGPA: 4.82/5.0 (First Class Honours)Class rank:1/87
Research Interests
- Communication Theory
- Wireless Communications
- Next-Generation Wireless Networks
- AI/ML
- Reinforcement Learning
Technical Skills
-
Python
-
C++
-
MATLAB
-
JAX
-
PyTorch
Projects
PyTwinNet
A modular Python library for simulating wireless digital twins and optimizing wireless networks.
Library Management System
A C# integrated library management system for monitoring books borrowed and patrons.
Realty Management App
An application for real estate companies.
Languages
-
Yoruba
-
English