I am a biomedical engineer and machine learning researcher working across medical imaging, embedded AI, and intelligent sensing. At Université Paris-Saclay and Lab-STICC, IMT Atlantique, I develop hardware-aware medical image segmentation and learned wireless communication systems through the FEF-AI4SD project. My work also spans assistive technologies, multimodal learning, and full-stack software engineering.
Kwame Nkrumah University of Science and Technology
GPA: 3.61/4.00. Thesis: Multi-Stage Deep Learning Framework for Thyroid Nodule Segmentation in Low-Resource Ultrasound Imaging.
Université Paris-Saclay, France · FEF-AI4SD • September 2026 - Present
Lab-STICC, IMT Atlantique, Brest, France · FEF-AI4SD • September 2026 - Present
Université Paris-Saclay, France · FEF-AI4SD • September 2026 - Present
Developed HALO-nnUAP, an automated hardware- and dataset-aware framework for 2D, 3D, and 4D medical image segmentation. Established multi-dataset evaluation, cross-validation, computational complexity analysis, GPU memory profiling, inference latency measurement, and MLflow-based experiment tracking. Developed computational spectroscopy models using FFT and power spectral density analysis. Supervisor: Professor Daniel Adjei.
Lab-STICC, IMT Atlantique, Brest, France · FEF-AI4SD • September 2026 - Present
Designed MLPARE-Com for joint physical- and link-layer optimization using learned transceiver models, hybrid channel estimation, and Q-learning-driven retransmission policies. Developed ESP32-S3 firmware for neural inference, real-time channel sensing, and adaptive ACK/NACK control, with an evaluation pipeline spanning MIMO simulation and embedded deployment. Supervisor: Professor Derek Asiedu.
Auphanim Technologies and Sciences · Kumasi, Ghana • June 2026 - Present
Designed an edge-deployed computer vision system that monitors RTSP/HLS camera feeds and autonomously escalates verified threats.
Mentismint · Accra, Ghana • October 2025 - Present
Built and shipped a React Native personal finance app with Expo, Expo Router, and EAS. Delivered full-stack features across a microservice backend, a typed TypeScript client/service layer, and a Groq AI integration.
Quantum Computing and Assistive Technologies Lab, KNUST • October 2025 - October 2026
Led Deafly Health, an AI-native multimodal emotion recognition engine for Ghanaian Sign Language communication. Designed the Spatio-Temporal Emotion Graph Network (ST-EGN), achieving 84.6% cross-modal emotion detection accuracy on preliminary datasets. Supervisor: Professor Jerry John Kponyo.
Responsible AI Lab, KNUST • October 2025 - October 2026
Contributed to MRI-CT cross-modality translation, implementing and evaluating deep learning methods for structurally faithful synthetic CT images in resource-limited settings. Supervisor: Professor Jerry John Kponyo.
Quantum Computing and Assistive Technologies Lab, KNUST • February 2026 - August 2026
Designed a dual-mode GQD-silicon nitride biosensor with a simulated IL-6 detection limit of 0.02 ng/mL. Developed 3-qubit IBM Qiskit models and simulated multiplexed PCT, CRP, and IL-6 sensing. Supervisor: Professor Jerry John Kponyo.
Bioinstrumentation Lab, KNUST • January 2025 - September 2025
Designed HALO-UNet for thyroid nodule segmentation, achieving a 71% Dice score on real-world ultrasound datasets. Developed a hardware-aware optimization pipeline with more than 70% faster inference under low-resource constraints, validated across datasets and architectural ablations, and deployed on Hugging Face Inference Endpoints. Supervisor: Dr. Isaac Acquah.
The Build Fellowship · Remote • March 2025 - May 2025
Optimized T cell differentiation using full factorial Design of Experiments. Developed a scalable suspension culture process with 18% lower production costs, over 90% cell viability, and batch variability reduced from 25% to 12%. Validated the protocol through three GMP-like runs. Supervisor: Claire Wang.
Masara Consult · Birmingham, United Kingdom • April 2024 - April 2025
Launched an AI-powered multi-curriculum e-learning platform reaching 1,000+ users in its first month. Built Node.js and SQL RESTful services for user management, course delivery, and analytics.
BUILD SCITECH · Kumasi, Ghana • March 2023 - Present
Leads product development across fintech, healthtech, agritech, and e-commerce. Established a technology expo community of 600+ African student developers, founders, and employers.
African German Entrepreneurship Academy · University of Leipzig • September 2023 - September 2025
Developed Agropremium, a web-based global trade platform for fair-trade logistics, and ObstaMed, a sensor-integrated assistive mobility cane. Supported by a €15,000+ travel grant.
IEEE Instrumentation and Measurement Society, KNUST Chapter • January 2024 - September 2025
Led mentorship and innovation programs, guiding students through instrumentation, embedded systems, and intelligent sensing workshops and conferences.
BUILD SCITECH • October 2023 - Present
Established a Technology Transfer Program and supervised over 10 student-led research projects in health, climate, biotechnology, and energy. Facilitated technical workshops on agentic system development.
ENACTUS KNUST • September 2023 - September 2024
Led the Projects & Innovation Division across 150+ members and seven projects, overseeing technical design and community pilots involving AI, IoT, and automation.
Biomedical Engineering Students’ Society, KNUST • September 2023 - September 2024
Led 10+ technical workshops in AI, IoT, and 3D printing for 200+ students, supporting academia-industry collaboration.
Oppong, R., Dotse, K., Adjei, A. A., Arthur, J., Kasore, V., Kponyo, A. B., Agyemang, J., Kponyo, J. J. KONVENS 2026, Hamburg, Germany. Accepted and presented as a poster. Hamburg/KONVENS 2025
[Paper]Oppong, R., Adjei G. S. iN4iN 2025, Leipzig, Germany. Oral presentation. Manuscript. Leipzig/iN4iN Conference
[Slide Deck]Oppong, R., Anokye, A. J., Acquah, I. TWEEK 2025, Kumasi, Ghana. Oral and poster presentation. Manuscript. Kumasi/TWEEK 2025
[Poster]Oppong, R. Preprint. October 2026. DOI: 10.20944/preprints202610.2580.v1.
[Preprint]
An integrated machine learning framework for physical- and link-layer optimization, combining variational autoencoder-based transceiver models, ML-based channel estimation, and reinforcement learning-based retransmission policies.
A hardware-aware segmentation framework spanning HALO-UNet, HALO-Former, and HALO-Temporal. Pilot HALO-UNet results across four CT/MRI datasets show up to 26.5% lower inference latency than nnU-Net.
A mobile ultrasound system built around HALO-UNet for thyroid diagnostics. Awarded an $8,000 pre-seed grant and the Mansa-to-Mansa NVIDIA Jetson Orin Nano project challenge ultimate prize.
A role-based peer learning marketplace with booking, scheduling, chat, student/tutor onboarding, and session management. Reached 1,000+ active beta users across one campus in four weeks.
An open-source, type-safe TypeScript request pipeline that models execution as a dependency graph. A topological scheduler enables parallel execution, circular dependency detection, and Next.js integration.
Python, Modern C++ (C++14/17/20), C, TypeScript, JavaScript, CMake, Async I/O, RAII
OpenCV, MONAI, SimpleITK, NiBabel, MedT, nnUNet, YOLOv26 Ultralytics
ESP32, ESP-IDF, FreeRTOS, Raspberry Pi, NVIDIA Jetson, ARM, Arduino, Embedded C/C++, HALs, ISRs
FFT, power spectral density, frequency-domain analysis
PyTorch, TensorFlow, ONNX, ONNX Runtime, NVIDIA TensorRT, ExecuTorch, TensorFlow Lite
NVIDIA CUDA, cuDNN, GPU profiling, CPU-GPU workload partitioning, ARM NEON SIMD, Tensor Cores, ONNX graph optimization
React, Next.js, React Native, Node.js, FastAPI, CI/CD, Docker, Unix/Linux, Git, Vim/Emacs/Nano, OOP, microservices
Probability and statistical inference, Bayesian estimation, computational complexity analysis, time and space complexity, error propagation, sensitivity analysis
English (Advanced)