Hello, I'm
Computer Science graduate from Islamic University of Madinah. Working on reinforcement learning, autonomous systems, and low-resource NLP.
I graduated in Computer Science from the Islamic University of Madinah, originally from the Central African Republic. My work focuses on reinforcement learning for robotic control, FPGA-based neural network inference, and building NLP tools for underserved languages.
Outside of academics, I have been involved in community development work in CAR, coordinating with international organizations on humanitarian initiatives.
Medina, Saudi Arabia
Islamic University of Madinah
Central African Republic
AI & Autonomous Systems
Developed a PPO-based autonomous navigation system for a Crazyflie quadrotor in the Webots simulator. Designed a hierarchical PID+DRL control architecture with curriculum learning across three difficulty stages. Conducted systematic reward engineering across seven iterations, demonstrating that positive reward gradients outperform punitive strategies for continuous robotic control. Ranked top graduation project of the department.
Paper under review at MDPI Drones • Certificate of Outstanding Achievement from the Faculty Dean
First dedicated NLP system for Sango, the national language of the Central African Republic spoken by approximately 5.5 million people. Constructed a quality-filtered parallel corpus of 21,125 Sango-French verse pairs, applied LoRA fine-tuning to NLLB-200, and built an integrated translation-plus-speech pipeline. To our knowledge, the first Sango-specific NLP system reported in the literature.
First dedicated Sango-French NLP system. Constructed a 21,125-pair parallel corpus and applied parameter-efficient fine-tuning to NLLB-200 using LoRA, achieving strong in-domain translation quality while preserving general-domain capability. Integrated with Meta's MMS-TTS for Sango speech synthesis. All resources publicly released.
PPO-based quadrotor navigation with obstacle avoidance in Webots simulation. Hierarchical PID+DRL architecture with curriculum learning. Systematic reward engineering across seven versions demonstrating the impact of reward polarity on continuous control. Ranked top graduation project. Certificate of Outstanding Achievement from the Faculty Dean.
320-hour intensive program at KAUST Academy. Built automated pipeline integrating AlphaFold2 and AutoDock Vina for computational biology. Focused on workflow optimization and reproducibility.
Configured MPI-based computing cluster for parallel computing coursework. Documented distributed systems setup and demonstrated parallel execution principles across networked nodes.
FPGA hex display system for the DE1-SoC (Cyclone V 5CSEMA5F31C6). Implements a reusable 7-segment decoder module supporting full hexadecimal (0–F) output across multiple displays using Verilog HDL.
Autonomous obstacle-avoidance robot powered by a neural network deployed entirely in FPGA hardware. The DE1-SoC (Cyclone V) runs inference in ~1 microsecond using a compact 3-8-3 network in Q6.10 fixed-point arithmetic. Zero collisions during live demonstration. Fully battery-powered and untethered.
Islamic University of Madinah, Saudi Arabia
5-year program including preparatory year • 136 total units
King Abdullah University of Science & Technology
320-hour intensive computational biology program. Selected participant.
Certificate of Outstanding Achievement — Graduation Project, Faculty of Computer & Information Systems (April 2026)
Recognition for contribution to ABET accreditation renewal process, Faculty of Computer & Information Systems
Participant — KAUST Academy 2025
Central African Republic
Réseau des Jeunes Centrafricains pour le Développement
Certificate of Recognition for community development involvement.
I'm always open to discussing research opportunities, collaborations, or just connecting. Feel free to reach out.
Your message has been received. I'll get back to you as soon as possible.