Murad Mebrahtu | Robotics Software Engineer

About Me

Robotics Software Engineer with expertise in developing and deploying AI/CV applications on edge devices for autonomous systems. Experienced in end-to-end development of perception algorithms and multi-sensor integration using ROS2, Python, and C++. Strong background in validating robotic systems through Gazebo and RViz simulation, with a focus on real-time performance. Proven track record in LiDAR data processing and sensor fusion for autonomous navigation in complex environments.

Technical Skills

Robotics & AI

ROS2, Gazebo simulation, RViz, sensor integration (ZED2 camera, Ouster LiDAR)
Real-time object detection and tracking, sensor fusion, multi-sensor calibration
Edge AI Deployment: ML model optimization, real-time inference
LiDAR Processing: Point cloud processing, 3D object detection, sensor fusion

Programming & Tools

Languages: Proficient in Python; Intermediate C++
ML/AI: PyTorch, CUDA, TensorFlow, transformer architectures for computer vision
Development: Git version control, modular software architecture
Simulation: Gazebo, RViz for algorithm validation

Languages

English: Very Good command (IELTS band 7)
Tigrigna and Bilen: Bilingual Proficiency

Work Experience

Research Engineer

Autonomous Vehicle's Lab (AVLab), Khalifa University

June 2023 - Present

Designed and implemented ROS2-based perception systems integrating ZED2 camera and Ouster LiDAR
Extensively used Gazebo and RViz to validate perception and control algorithms
Developed and deployed computer vision and AI applications on edge devices
Designed modular perception systems that seamlessly integrated with planning modules
Collaborated on designing modular, scalable stack architecture for autonomous vehicle software
Led annotation validation, script development, and benchmark testing for proprietary datasets
Represented Khalifa University in the Abu Dhabi Autonomous Racing League (A2RL) in 2024
Showcased robotics and autonomous systems innovations at prestigious conferences

Graduate Teaching and Research Assistant

Khalifa University

January 2021 - May 2023

Developed probabilistic trajectory prediction algorithms for autonomous navigation
Served as Lab Instructor for Programming Concepts, MATLAB, and C++ courses

Software Developer (Freelance)

GamaLearn

December 2020 - February 2021

Developed automated recruiting system with candidate assessment and qualification pipeline

Software Developer Intern

GamaLearn

June 2020 - August 2020

Created Business Card Scanner application using Power Apps and Power Automate for automated database integration

Education

MSc. in Computer Science

Khalifa University, Abu Dhabi, UAE

January 2021 - May 2023

GPA: 3.93 / 4.0

Thesis: Probabilistic Pedestrian Trajectory Prediction For Autonomous Vehicles

Focus: Computer Vision, Machine Learning, Robotics, Autonomous Systems

BSc. Computer Engineering

Khalifa University, Abu Dhabi, UAE

January 2016 - December 2020

GPA: 3.8 / 4.0 (Excellent With Highest Honor)

Senior Design Project: MultiModal Biometric Authentication using Machine Learning

High School Diploma

St. Joseph's School, Keren, Eritrea

September 2010 - September 2014

GPA: 3.8 / 4.0

Projects & Research

Autonomous Vehicle Perception System

2023

Developed end-to-end perception pipeline integrating LiDAR and camera data
Implemented and optimized detection algorithms (YOLOv10, PointPillars) for edge deployment
Designed sensor synchronization protocols and calibration procedures

Probabilistic Pedestrian Trajectory Prediction

2021-2023

Designed transformer-based prediction model for real-time pedestrian trajectory forecasting
Implemented ML pipeline with PyTorch and CUDA optimization for edge device deployment
Integrated prediction module with autonomous vehicle planning system

LiDAR Processing and Multi-Sensor Fusion

2022-2023

Implemented point cloud processing algorithms for LiDAR data
Developed sensor fusion techniques combining LiDAR with camera data
Optimized LiDAR processing pipeline for real-time performance

Parasitic Egg Detection and Classification

2021

Implemented computer vision techniques for automated detection and classification
Applied machine learning algorithms to improve diagnostic accuracy in medical imaging

MultiModal Biometric Authentication

2020

Designed and implemented ML-based authentication system using multiple biometric inputs
Developed multi-factor security protocol with improved accuracy and security features

Publications & Presentations

Towards Autonomous and Safe Last-mile Deliveries with AI-augmented Self-driving Delivery Robots

E Shaklab, A Karapetyan, A Sharma, M Mebrahtu, M Basri, M Nagy, ...

arXiv preprint arXiv:2305.17705 (2023)

10 citations

Transformer-Based Multi-Modal Probabilistic Pedestrian Prediction For Risk-Aware Autonomous Vehicle Navigation

M Mebrahtu, A Araia, A Ghebreslasie, J Dias, M Khonji

2023 21st International Conference on Advanced Robotics (ICAR), 652-659 (2023)

1 citation

EMT: A Visual Multi-Task Benchmark Dataset for Autonomous Driving in the Arab Gulf Region

NA Madjid, M Mebrahtu, A Nasser, B Hassan, N Werghi, J Dias, M Khonji

arXiv preprint arXiv:2502.19260 (2025)

Probabilistic Pedestrian Trajectory Prediction For Autonomous Vehicles In Crowded Environments

MS Mebrahtu

Khalifa University of Science (2023) - Master's Thesis

Conference Presentations

Presenter, International Conference on Advanced Robotics (ICAR)
Exhibitor, International Conference on Intelligent Robots and Systems (IROS)
Exhibitor, Intelligent Transportation Systems Conference (ITS)
Exhibitor, Unmanned Systems Exhibition (UMEX) and SimTEX