I build systems that connect AI with the physical world: cameras, sensors, microcontrollers, robots, local servers, and real users.

My work focuses on edge AI, embedded systems, robotics, and learning technologies — especially in contexts where cloud access, connectivity, hardware resources, and deployment conditions cannot be taken for granted.

🔌 Edge AI & Embedded ML Offline-first systems, on-device inference, computer vision, TinyML — Raspberry Pi, ESP32, RFID, sensors, and local-first architectures.

🤖 Robotics & Human-Machine Interaction Voice-controlled robots, educational robotics, robotic arms, semantic command interfaces, and AI-assisted control systems.

📚 Learning & Productivity Technologies Tools that help people learn, focus, build skills, and access technology more effectively — in African and resource-constrained contexts.

Offline-first biometric attendance that runs on-device — face + RFID, no cloud required.

A field-oriented AIoT attendance system designed for environments with unreliable connectivity and limited infrastructure.

Python · FastAPI · React · Raspberry Pi · InsightFace · ChromaDB · MQTT · PostgreSQL

Indoor-following cart prototype using BLE localization and embedded control.

C++ · ESP32 · BLE · MQTT · Python · Arduino

Voice-controlled educational robot framework using speech recognition, semantic command matching, vector search, and native robot control.

Python · Speech Recognition · Embeddings · ChromaDB · tdmclient

Incident-detection prototype for controlled environments.

Python · Streamlit · Computer Vision

Privacy self-audit assistant for public digital footprint awareness.

Python · AI Agents · LLMs · Streamlit

I am consolidating my work around one core idea:

Build intelligent systems that remain useful when the hardware is limited, the cloud is unavailable, and the problem is real.

My long-term direction combines edge AI, robotics, privacy-aware computer vision, African language technologies, and AI tools for learning and productivity.

• 🔧 Build for real hardware, not only notebooks.
• 🔒 Prefer offline-first and privacy-aware architectures.
• 📉 Design for constrained environments.
• 👥 Keep systems understandable for real users.
• 🧠 Treat AI as part of a complete system, not a standalone trick.

Languages

Embedded & Edge

AI & Vision

Robotics

Backend

Frontend

Some of my prototypes explore sensitive areas such as computer vision, public digital footprint analysis, and automated decision support.

These projects are designed as research, educational, and self-assessment tools. They are not intended for unauthorized profiling, harassment, doxxing, stalking, identity tracking, or biometric identification without consent.