Gedeon Muhawenayo



Cinque Terre

Gedeon Muhawenayo
gedeonmuhawenayo@gmail.com

I am a machine learning research engineer at INRIA centre Grenoble-Rhone-Alpes. I have a masters degree in Machine Intelligence from the African Masters in Machine Intelligence (AMMI), where I worked with Georgia Gkioxari. I am interested in Computer Vision (satellite and aerial images processing), Graph Convolution Network and Model Optimization (Pruning, Quantization and Binarization). I have a Bachelors degree in Electronics and Telecommunication Engineering from the University of Rwanda.


Cinque Terre

Msc. Machine Learning.
Oct 2019 - Nov 2020

I completed an intensive Machine Learning program at the African Masters in Machine Intelligence(AMMI). A Master’s program founded by Google and Facebook which provides a state-of-the-art training in Machine Learning and its Applications. Here are the courses I have covered and grades I have obtained so far.
Cinque Terre

Bsc. Electronics and Telecommunication Engineering.
Sept 2015 - June 2019

I did my Bachelors degree in Electronics and Telecommunication Engineering at University of Rwanda where I graduated with a cumulative average of 80.09%. Here are the modules I covered and grades I obtained.

game

Compressed Object Detection
In this work, we extended pruning, a compression technique which discards unnecessary model connections, and weight sharing techniques for the task of object detection. With our approach we are able to compress a state-of-the-art object detection model by 30.0% without a loss in performance. We also show that our compressed model can be easily initialized with existing pre-trained weights, and thus is able to fully utilize published state-of-the-art model zoos.


game

Cassava Disease Classification.
The project involved classifying pictures of cassava leaves into 1 of 4 disease categories (or healthy). The disease categories included: Cassava Mosaic Disease(CMD), Cassava Brown Streak Disease(CBSD), Cassava Bacterial Blight(CBB), and Cassava Green Mite(CGM).


game

Object Tracking.
This project was about detecting in time series data. The idea was to track objects in a video, so it was to compare the detected instances in two consecutive frames within a series of frames that make the video. My approach was to use color coding and track the number of detected objects.


game

Common Image Operations with OpenCV.
This project was about using openCV for common image operations. The overall operations covered include; Basic image operations on image, Arithmetic operations on image, Code performance and improvement in openCV, Image thresholding, Geometric transformation on an image, Smoothing an image, Morphological transformation of an image, ImageGradient, CannyEdgeDetection and ImagePyramids


game

Binary Graph convolutional Neural Networks.
The goal of this project is to implement an efficient graph CNN by using binary weight, This means all operations such as matrix multiplication will be done by using a bitwise operation. As the project is still under development and the paper is not yet submitted and for this reason the implementation is currently private. But here is the implementation of the initial work of the overall work which is graph classification.


game

Signal and Functions Analyzer UI.
I designed a windows Graphical User Interface (GUI) which enables students to do most of their simulations and to do live plots of electric signals and mathematical functions. The design was inspired by Mathematics, digital signals processing and UI design. The implementations were done using Python and C++.


game

Object Detection, Segmentation and Human pose estimation.
This project was kind of reproducing the Mask RCNN paper basically is about doing instance segmentation and object detection. I did my implementations with Detecron2 instance segmentation pretrained model as well as pose estimation whose backbone is Resnet50.


game

Smart Electrical Energy Meter
This is a device to allow users to share one electrical energy meter, the design optmized cost and system quality particularly power, voltage and current ratings. The project is designed by incorpolating microprocessor skills, electrical&electronics engineering and software development


game

Attendance system
Attendance system was designed and implemeted by a team of seven students at University of Rwanda , The main idea was to create an efficient way of collecting student and lectures attendance records without relying on internet connectivity. Furtermore, the device is portable and it runs on battery, which makes it suitable in rural areas.


game

Embedded system and IoT
These are projects that I did in my undergrad studies, most of them are implemeted by a combination of Python and C++. integreting sensor and rapid prototyping boards(Arduino, Raspberry pi), these are projects that I designed and implemented; Automatic coocker, Car overturn accident alert via GSM, Heartbeat and body temperature monitoring system, Blind stick, Irrigation system, school_bell, acqualium_monitoring, coins_based_juice_vending_machine, solar_powered_irrigation_system, speak_for_dumb, pills_dispenser, fish_farming, attendance_system, boat_accident_alerting_system




Certificates

  • High Performance Computing (HPC): Trained by PSSC Labs on PowerWulf ZRX1+ HPC Cluster and CBeST trainin. Covered HPC cluster installation, Testing and operation. The cluster has CPU nodes, GPU node and Hadoop, With Intel Omnipath® and Mellanox® High Speed Network Interconnects.


Computer programming and Frameworks

  • Python: I am familiar with openCV, Pytorch , Scikit learn , Detectron2, numPy and pandas among others.
  • C and C++: I have experience with low level programming and memory constrained systems programming.
  • MATLAB: I have good understanding of MATLAB for basic image processing operations and signals processing.
  • GCP: I am familiar with Google Cloud Platform(GCP) particularly cloud GPUs.


Internships

  • Computer vision reseach Internship: From March 2020 up to now, I am doing a remote based intership at Spectrum AI where am working with Rajat Mani Thomas on binary graph convolutional neural Networks.
  • Software Developmet intership: I did a Lab based software development intership at Hview tech for a period of 3 months.
  • IoT and Embedded systems Intership: I have worked at Fablab Rwanda an intership in the field of IoT and embedded systems, under the supervision of researchers from Kobe Institute of Computing for 6 months.



game

Conference on Computer Vision and Pattern Recognition (CVPR)
I was among few young researchers volunteered at CVPR2020, It has helped me to contribute in the arangement and testing conferencing Platform, it also allowed me to attend the conference where I got access to some of the state of the art in computer vision.


game

Machine Learning Instructor at 10Academy Data Science Training
I volunteered at 10academy, as an instructor in the Data Science training. I prepared and delivered the tricks for winning a data science competition and challange. The notes on data challenges can be found here. And the notebook on Multilayer Perceptron can be found here.