Research
My research uses signal processing and machine learning to solve practical problems: keeping personal data safe in AI systems, making wireless links more reliable, and building low-cost systems that work in places with little infrastructure. I work with faculty in the Engineering department at Ashesi University.
Current Work
Signal-Level Privacy for Machine Learning
Can data be protected before it ever reaches a model? With Prof. Hyder Ali Segu Mohamed, I am testing input-first quantization as a way to limit what a trained model leaks about its training data, while keeping it accurate.
Neural Channel Prediction for OTFS
Context-aware neural prediction of FIR filter parameters for OTFS-based wireless channels that change over time. Presented at IEEE ICAST 2026.
Hardware Acceleration for Imaging
An FPGA-based FFT accelerator for detecting blur and restoring images in UAV surveillance. Co-authored with S. Agudu and S. Armah; presented at INTCEC 2026.
Edge AI for Agriculture
A YOLOv8 detector running fully offline on a Raspberry Pi 4 to keep birds off rice farms, with 98% detection accuracy. Presented at BCUR 2026 in Glasgow.
Earlier Research
-
Oct 2024 – Mar 2025 · African Leadership AcademyAI for Health Research FellowshipAssessed the quality of treated drinking water in communities affected by illegal mining (galamsey) in Ghana.
-
Jan – Aug 2024 · Ashesi UniversityML-Based Electronic Scarecrow (BSc thesis)
-
May – Jul 2024Cooperative Spectrum Sensing in Cognitive RadioImplemented cooperative sensing algorithms in MATLAB and Python and simulated how well they detect free channels.
-
2023 – 2024Power Systems: Changeover Switch & Microgrid ControlA three-phase automatic changeover switch, published in the Ashesi SEED journal, and a study of hierarchical control for microgrids with renewable sources.
-
Nov – Dec 2021Optimization Using the q-Gradient MethodStudied how Jackson's derivative helps gradient search escape local minima, and compared it with genetic algorithms.