CV
CV of Mohammed, robotics engineer in SLAM, navigation and computer vision. The PDF is behind the icon on the right.
Contact Information
| Name | Mohammed |
| mohammed.gsr-5161-16@aau.edu.et | |
| Website | https://berito.github.io |
Experience
-
2026 - present Ethiopia
Part-time Lecturer, Robotics
Addis Ababa University
- Introduction to Robotics and Intelligent Systems (fourth-year BSc Software Engineering): lectures and weekly labs, from robot motion and control and sensing to Kalman filtering, localization and mapping, and path planning, run on a simulated mobile robot in ROS 2 and Gazebo.
- Robotics and Intelligent Systems (MSc in Artificial Intelligence), lab assistant: ROS 2 labs on simulation, perception, sensor fusion and SLAM, and robot arm motion planning with MoveIt 2.
- Lab material published on GitHub.
-
2025 - present Freelance Robotics Engineer
Upwork (remote)
- Unitree L2 LiDAR SLAM (solo): selected and integrated the SLAM stack for the client’s solid-state LiDAR (Point-LIO odometry with a GTSAM pose-graph back-end), set up its UDP driver and IMU, and handed over in Docker.
- Jetacker Nav2 tuning (solo): built the simulation and Nav2 stack as about 20 Docker containers and tuned planners, costmaps and obstacle avoidance with a structured method to 30 of 30 runs on the walls map and 26 of 30 with obstacles; patched a
use_sim_timedeadlock inrobot_localizationfrom source; reported an integrity problem in the client’s data. - SkyHunter multi-robot UGV simulation (team; my parts: simulation, communication, localization and navigation): built the Gazebo simulation from the robot’s CAD (STEP) files, with LiDAR, IMU, GPS and gimbal cameras and spawning of up to 8 robots; set up EKF localization, GPS in the local frame with
navsat_transform, Nav2 and a 27-frame TF tree; wrote the communication package, a failover state machine across WiFi, LTE and LoRa with measured latency. - erthbot thermal inspection (team): built the UI of a deployed solar-panel thermal inspection system.
- SLAM tutoring (1-on-1): pose-graph optimization, factor graphs and LiDAR-inertial odometry.
-
2021 - present Ethiopia
Assistant Lecturer, Electrical and Computer Engineering
Samara University
- Teach undergraduate courses in C++ programming, Python, operating systems, relational database systems and VLSI design.
-
2020 - 2023 Ethiopia
Co-Founder (part-time)
Elamiso Consulting
- Software and IT consultancy: ERP implementation on Odoo and software deployment.
-
2021 - 2023 Ethiopia
Instructor and Coordinator
High-School STEM Program, Samara University
- Coordinated and taught a weekend STEM program across two cohorts: Arduino, electronics, programming, 3D printing and wheeled robots, ending in team capstone projects.
- Mentored student teams that won 1st place at a national innovation competition (2023).
-
2018 - 2021 Ethiopia
Software Developer
ICT Business Application Development, Samara University
- Developed the university’s business applications in Java and Spring Boot (microservices, PostgreSQL).
Summary
- Robotics engineer building autonomous ground robots and drones in ROS 2 and simulation. Tuned Nav2 for a client robot to 30 of 30 benchmark runs, built multi-robot Gazebo simulation for up to 8 robots, and flew PX4 offboard laps that repeat within 6 cm.
Projects
-
in progress MSc thesis: robust pose-graph optimization
C++17, GTSAM, g2o
- Correlation-aware robust pose-graph optimization; baseline campaign of robust back-ends (GNC, DCS, Switchable Constraints) with my own precision, recall, ATE and RPE evaluator.
-
PX4 track pacing drone
PX4 SITL, ROS 2, MAVROS, Gazebo Harmonic, Python
- A drone that flies a configured route lap after lap: offboard flight node with a state machine, track geometry with velocity and acceleration feed-forward, and a Gazebo world generated from the same geometry. Laps 2 and 3 within 6 cm of each other (PX4 flight log). Video on YouTube.
-
Mengabot multi-robot simulation
ROS 2, Gazebo Harmonic
- A modular multi-robot ground platform in simulation, reusable for any ground robot, running up to 8 robots. Video on YouTube.
-
Mengabot camera perception
C++, ONNX Runtime, ROS 2, Gazebo Harmonic
- YOLOv8 detection, ByteTrack tracking and a Kalman-filter target estimator with lead point. On a 0.275 m target crossing at 0.5 m/s, 5 m away: detected in 194 of 194 images, tracked in 133 of 137 messages, position median error 0.19 m, velocity 0.49 m/s against 0.5, against simulation ground truth.
Education
-
2023 - 2026 Ethiopia
MSc
Addis Ababa University
Artificial Intelligence
- Expected December 2026. Thesis in progress: robust pose-graph optimization for SLAM.
- Coursework: Artificial Intelligence, Machine Learning, Deep Learning, Computer Vision, Reinforcement Learning, Probabilistic Graphical Models, Natural Language Processing, Distributed Systems for AI, Cognitive Science.
-
Eritrea
BSc
Eritrean Institute of Technology
Computer Engineering
Technical Skills
State estimation and sensor fusion: EKF (robot_localization), particle filters, factor graphs and pose-graph optimization (GTSAM, g2o), robust estimation (M-estimators, GNC, IRLS, DCS, Switchable Constraints), Lie groups SE(3), IMU processing and time sync, GNSS with navsat_transform
SLAM and mapping: LiDAR-inertial SLAM (Point-LIO), loop closure, globally consistent mapping, TF chains, geographic frames (WGS84, ENU, UTM), evaluation (ATE, RPE)
Navigation: Nav2 tuning and benchmarking (costmaps, controller, planner), MCL/AMCL
Simulation: Gazebo Harmonic, multi-robot simulation, URDF/Xacro, CAD to robot model, wireless link simulation
Drones: PX4 SITL, offboard control with MAVROS, ArduPilot SITL
Perception: YOLOv8 in ONNX Runtime (C++), ByteTrack, Kalman-filter target estimation, PCL
Programming and tools: C++17, Python, ROS 2, PyTorch, Docker, Git, Linux, FastAPI
Teaching and technical writing: University teaching since 2021, lab material, client technical reports
Achievements
- Student teams I mentored and coordinated won 1st place at a national innovation competition, 2023.