// from firmware to full autonomy

Ahmed Mady
Robotics & Embedded
Software Engineer

Stuttgart, Baden-Württemberg, Germany
> ROS 2 · MoveIt 2 · Embedded C++ · SLAM · Navigation · 3D Localization · Fleet Management █
C/C++ Python ROS 2 MoveIt 2 Nav2 SLAM STM32 Jetson Orin FreeRTOS Isaac Sim
→ View Projects Get in Touch
Ahmed Mady
About Me

Where Hardware
Meets Intelligence

I'm a Robotics & Embedded Software Engineer based in Stuttgart, Baden-Württemberg, Germany, working on the 4NE-1 humanoid robot at Neura Robotics. I recently completed my MSc in Robotics and Artificial Intelligence at FAU Erlangen-Nürnberg, specializing in Planning & Control and Human-Machine Interfaces.

My work spans humanoid manipulation, imitation learning, and sim-to-real transfer — from collecting tele-operation data in the Neura Gym and fine-tuning NVIDIA Isaac GR00T models, to bridging simulation and hardware with Apple Vision Pro in Isaac Sim NeuraVerse.

I bridge the gap between bare-metal embedded systems and high-level autonomy stacks, building systems that move, sense, and reason in the real world.

MSc. Robotics & AI
FAU Erlangen-Nürnberg
Apr 2024 — Apr 2026 · Erlangen, Germany
Thesis: Multimodal Reasoning & Explainability for Robotic Manipulation in Physical AI
BSc. Mechatronics & Automation
German University in Cairo (GUC)
Sep 2018 — Jun 2023 · Cairo, Egypt
Thesis: Feedback Control System for a Series Elastic Actuated Robot
5
Companies
10+
Projects
45
GitHub Repos
2
Degrees
Fleet Management
Computer Vision
Poppy Robotic Arm
Experience

Where I've Built Things

Neura Robotics
Robotics Engineer — PoC Team, 4NE-1 Humanoid Robot
Jun 2026 — Present · Metzingen, Germany
  • Collected training data in the Neura Gym of the 4NE-1 humanoid robot performing household tasks (sorting clothes, loading/unloading washing machines and dryers) from tele-operation to train on using imitation learning for BSH client project.
  • Fine-tuned NVIDIA Isaac GR00T models using the recorded data to create a foundation perception-language-action (PLA) model.
  • Recorded tele-operation data in simulation using Apple Vision Pro within the Isaac Sim NeuraVerse environment, generating synthetic demonstrations for GR00T fine-tuning to replicate the simulated setup in hardware, bridging the sim-to-real gap for humanoid manipulation tasks.
4NE-1 Isaac GR00T Imitation Learning Isaac Sim Apple Vision Pro Sim-to-Real
Siemens AG
Robotics Working Student — Autonomous Systems
Jul 2025 — May 2026 · Munich, Germany
  • Built a robot fleet control setup for humanoid G1 robots and AGVs, coordinating tasks using VDA 5050 and Siemens SIMOVE
  • Set up multi-robot fleet management with OpenRMF, connecting it to ROS 2 navigation for distributed operation
  • Implemented navigation logic in C++ as part of the Siemens Closed Loop Intelligence Framework (CLIF)
  • Tested and tuned autonomy behavior on real robots, iterating based on sensor noise, localization drift, and navigation failures
  • Tested and compared 3D localization methods (MOLA, RTAB-Map, GLIM, KISS-SLAM, FAST-LIO) on humanoid robots using 3D sensor data, analyzing performance with defined KPIs
ROS 2 C++ SIMOVE OpenRMF VDA 5050 SLAM Nav2
Fraunhofer IGCV
Research Assistant — Robotic Manipulation
May 2025 — May 2026 · Augsburg, Germany
  • Used NVIDIA Cosmos multimodal LLM to plan manipulation tasks from natural-language commands for a UR5e robot
  • Built a language-to-motion pipeline, converting text instructions into waypoints and executable actions
  • Implemented UR5e manipulation in Isaac Sim using MoveIt 2, including pick-and-place and robotic screwing operations
  • Compared Cosmos-based control with reinforcement learning baselines, evaluating task success and explainability using LLM- based judging
UR5e MoveIt 2 Isaac Sim NVIDIA Cosmos Reason2 NLP C++
Lemvos Robotics
Embedded / Robotics Software Intern
Apr 2025 — Jun 2025 · Augsburg, Germany
  • Developed a ROS 2 control node to interface the embedded steering and propulsion system with higher-level control software
  • Tested embedded sensors on an Automate Pro device running ROS 2 in Docker on Jetson Orin, debugging sensor data and system behavior on hardware
  • Worked on embedded C++ software for marine steering and propulsion, implementing CANopen communication for motor control and actuator feedback
C++ CANopen ROS 2 Control Jetson Orin Docker
Robotics Corner
Embedded / Robotics Software Engineer
Jul 2023 — Apr 2024 · Cairo, Egypt
  • Taught C++ and basic software engineering and supervised embedded systems and robotics projects for client work
  • Developed embedded drivers in C/ C++ for AVR and ARM microcontrollers, utilizing GPIO, PWM, communication, and RTOS
  • Built control and motion planning software for a 6-DOF industrial robot arm using MoveIt 2, Gazebo, URDF, and ROS 2 control
  • Built software for an autonomous warehouse vehicle, using ROS 2 navigation, sensor data from LiDAR, IMU, and odometry, and SLAM for localization
C/C++ ROS 2 MoveIt 2 Gazebo STM32 AVR FreeRTOS SLAM
Technical Arsenal

Skills & Technologies

⚡ Programming & Tools

C++ (17/20) Python C Bash Git CMake Make Docker OOP GitLab CI

🤖 Robotics & ROS

ROS 2 MoveIt 2 Nav2 URDF Gazebo RViz SLAM AMCL EKF MPC Behavior Trees ROS 2 Control Isaac ROS

🔧 Embedded Systems

ARM Cortex-M/A STM32 Jetson Orin Raspberry Pi FreeRTOS Yocto Linux CANopen

🧠 AI & Machine Learning

PyTorch TensorFlow Keras RL / RLHF CNNs YOLO U-Net Transformers
CONTROL
Computer Vision
COMPUTER VISION
MODEL PREDICTIVE CONTROL
OPEN LOOP CONTROL
Featured Work

Projects

Click any project for details and GitHub link.

Credentials

Certifications

Blog

Thoughts & Writeups

Contact

Let's Build Something

Always open to discussing new projects, robotics challenges, or opportunities in embedded systems and autonomous technology.

madyahmad19@gmail.com ahmadmadyy ahmed-mady-am19