01 / Overview

Estimation you can fly on.

I build the perception and state-estimation systems that let autonomous machines know where they are — when GPS cannot tell them.

8+
Years in perception & autonomy
7
Peer-reviewed publications
5
Open-source releases
4
Countries worked in

Principal SLAM and computer-vision engineer with 8+ years delivering production autonomy software across UAV, automotive and defence programmes — from the maths of the estimator down to the NPU it runs on.

My work sits where sensing meets certainty: visual-inertial odometry, simultaneous localisation and mapping, and factor-graph optimisation over camera, LiDAR, IMU, GNSS/RTK and event streams. I own full system architecture and the autonomy release train — SITL/HIL verification, CI/CD, reproducible platform images — and I ship it to embedded and edge-AI hardware where the latency budget is measured in milliseconds.

I report to CTO level, interface directly with customers and drone OEMs, lead technical hiring, and publish in IEEE venues as an active IEEE TIM reviewer.

What I work on

  • 01

    State estimation & SLAM

    Visual-inertial odometry, EKF/ES-EKF, factor-graph and graph-based optimisation with GTSAM, Ceres and g2o.

  • 02

    Multi-sensor fusion

    Camera, LiDAR, IMU, GNSS/RTK and event cameras — with online, offline and mass-production calibration.

  • 03

    GNSS-denied navigation

    Drift-free global localisation for platforms operating where satellite navigation is unavailable or unreliable.

  • 04

    Perception at the edge

    Real-time deep perception on Jetson, Hailo-8 and Blaize NPUs — TensorRT/CUDA and ONNX Runtime optimisation.

EU-authorised to work — Netherlands, Highly Skilled Migrant.

02 / Experience

Eight years, one trajectory.

Each role is a keyframe on the path. Hover one to find it in the map.

  1. Jan 2026Present

    Principal SLAM & Computer Vision Engineer

    KRUSH Labs B.V. — Autonomy Team · High Tech Campus, Eindhoven, NL

    • Own the full autonomy system architecture for GPS-denied flight — from sensor selection and compute budgeting through state estimation, perception and control interfaces — reporting to the CTO on technical strategy and roadmap.
    • Lead the visual-inertial odometry and absolute visual navigation programme, delivering drift-free global geo-localisation for flight in GNSS-denied conditions.
    • Manage the autonomy release train: versioning strategy, CI/CD, reproducible build images, and coordinated integration across firmware, perception and ground-control software.
    • Established the software- and hardware-in-the-loop verification and validation loop used for regression testing ahead of flight campaigns.
    • Deploy real-time perception to embedded edge hardware in C/C++, Python and ROS 2, including model optimisation and camera–IMU calibration for production.
    • Act as technical interface to customers and drone OEMs, and lead technical hiring for the autonomy team.
    • System architecture
    • SLAM
    • Visual-inertial odometry
    • Sensor fusion
    • ROS 2
    • Embedded C++

  2. Feb 2025Jan 2026

    Postdoctoral Researcher

    University of Twente — ITC Faculty · Enschede, NL

    • Off-road SLAM and advanced perception in collaboration with the Dutch Ministry of Defence and Demcon; co-supervised interns and MSc thesis projects.
    • Led a robust LiDAR/camera/IMU sensor-fusion architecture for semantic segmentation in forestry environments.
    • Implemented deep-learning perception in MATLAB/Python, C/C++, PyTorch, ROS 2 and ONNX Runtime, with model optimisation for CUDA and TensorRT execution providers.
    • Off-road SLAM
    • LiDAR fusion
    • Semantic segmentation
    • PyTorch
    • TensorRT

  3. May 2023Dec 2024

    Algorithm Engineer

    Continental Engineering Services · Burgess Hill, UK

    • Delivered MLC50/55/60 AUTOSAR C++14 compliant software for an automated parking solution in production at a British OEM.
    • Developed IMU/RTK-GPS optimisation-based sensor fusion for a drone odometry pipeline on eCAL middleware.
    • Created genetic-algorithm synchronisation for 12 ultrasonic sensors, validated by Monte Carlo simulation.
    • Designed a Qt5/C++ plugin for SiL and HiL testing of an SR-Radar parking solution, and integrated SRR630/ultrasonic fusion in embedded C++ for software-defined vehicles.
    • Built collision-free USS firing-pattern hashing with Gaussian Mixture Models and MD5, applied across 10K+ vehicles; and a transformer-based method generating firing sequences on-vehicle in real time.
    • Delivered technical presentations and proposals to major British, Indian and American automotive OEMs.
    • AUTOSAR C++14
    • eCAL
    • Qt5
    • SiL / HiL
    • ISO 26262

  4. Oct 2020Oct 2023

    Doctoral Researcher

    IBISC Laboratory — SIAM Team, Université Paris-Saclay · Essonne, FR

    • Researched multi-sensor fusion across RGB-D and event cameras with IMU/GPS for autonomous navigation.
    • Taught C/C++ programming and electronic circuits, and supervised master's projects.
    • Served as PhD students' representative and organised the laboratory's annual PhD day.
    • Event cameras
    • RGB-D
    • Teaching

  5. Feb 2020Sep 2020

    Research Assistant — Sensor Fusion

    IBISC Laboratory — MALIN Challenge · Essonne, FR

    • Implemented ES-EKF multi-sensor (IMU/GPS-camera) fusion for a localisation system.

  6. Feb 2019Aug 2019

    Teaching Assistant

    Nile University — Computer Engineering · Giza, EG

    • Taught Computer Systems Software, Logic Design, Embedded Systems and Computer Security.

  7. Jul 2018Nov 2018

    Embedded Systems Intern

    EmbeddedFab · Giza, EG

    • Developed embedded software and drivers for industrial sensors and actuators; participated in the Valeo Testing Academy.

Education & credentials

  • Feb 2024

    Lecturer Qualification

    French Ministry of Higher Education and Research

    Section 61 — Computer Engineering, Automation & Signal Processing · No. 24261392899

  • Oct 2023

    Ph.D. in Signal & Image Processing

    Université Paris-Saclay, France

    Thesis: Visual Odometry Using Heterogeneous Cameras for Simultaneous Localization and Mapping for Autonomous Vehicles

  • Sep 2020

    M.Sc. Smart Aerospace & Autonomous Systems

    Université Paris-Saclay, France

    Grade 16.778/20 · Ranked 1st

  • Jun 2018

    B.Sc. Aerospace Engineering

    Cairo University, Egypt

    Grade 88.89% · Honours

  • 2023

    Sensor Fusion Engineer Nano-Degree

    Udacity

    Certification

03 / Research

Published work.

Seven peer-reviewed papers and a doctoral thesis. Full record on Google Scholar.

  1. 2026

    Towards Autonomous Off-road Navigation in Forests

    Workshop ICRA 2026

  2. 2024

    DH-PTAM: A Deep Hybrid Stereo Events-Frames Parallel Tracking and Mapping System

    Journal IEEE Transactions on Intelligent Vehicles

    First author — conception, implementation, evaluation.

    Paper Code

  3. 2024

    GPS-Enhanced RGB-D-IMU Calibration for Accurate Pose Estimation

    Journal Communications in Computer and Information Science, Springer

    First author — theoretical conception, implementation, evaluation.

    Paper

  4. 2023

    Visual Odometry Using Heterogeneous Cameras for Simultaneous Localization and Mapping for Autonomous Vehicles

    Ph.D. Thesis Université Paris-Saclay · NNT 2023UPAST119

    Defended 5 October 2023.

    Thesis

  5. 2023

    IBISCape: A Simulated Benchmark for Multi-modal SLAM Systems Evaluation in Large-scale Dynamic Environments

    Journal Journal of Intelligent & Robotic Systems

    First author — conception, implementation, evaluation.

    Paper Code

  6. 2023

    MAV Localization in Large-Scale Environments: A Decoupled Optimization/Filtering Approach

    Journal Sensors — MDPI

    First author — conception, implementation, evaluation.

    Paper VIO_RGB_IMU B-splines

  7. 2023

    Robust RGB-D-IMU Calibration Method Applied to GPS-Aided Pose Estimation

    Conference VISAPP 2023 · Lisbon, Portugal

    First author — conception, implementation, evaluation.

    Code

  8. 2023

    Flow-Based Visual-Inertial Odometry for Neuromorphic Vision Sensors

    Conference VISAPP 2023 · Lisbon, Portugal

    Contributing author — C++ optimisation software, extensive evaluation.

  9. 2022

    Towards an Event-based Colour-encoded Vision for Robotics

    Invited talk Emerging Visual Sensors for Robotics, GdR ISIS · Sorbonne University, Paris

    Programme

  10. 2022

    Heterogeneous Sensors — Fusion and Odometry

    Talk IBISC PhD Students' Annual Day · Pelvoux, France

    Programme

04 / Projects

Built and shipped.

Lead Algorithm Engineer · AEROINTEL B.V., Amsterdam

Drone Odometry Pipeline

An embedded AI perception stack delivering real-time, on-board sensor fusion for autonomous-flight research and development.

  • Thread-safe pipelines with separate producer threads for serial I/O and camera capture, holding a real-time budget on embedded hardware.
  • Extended Protobuf schema over MAVLink telemetry for structured, language-agnostic exchange.
  • Camera–IMU fusion for on-board odometry in C++.
  • Map-aided visual localisation to supplement inertial navigation.
  • Embedded C++
  • Sensor fusion
  • Protobuf
  • MAVLink
  • Edge AI

Open source

  • DH-PTAM

    Deep hybrid stereo events-frames parallel tracking and mapping — combining learned front-ends with classical SLAM for robustness in hard lighting and motion.

    IEEE T-IV 2024

    Repository

  • IBISCape

    A CARLA-based simulated benchmark for evaluating multi-modal SLAM in large-scale dynamic environments, with ground-truth trajectories and calibration sequences.

    J. Intell. Robot. Syst. 2023

    Repository

  • VIO_RGB_IMU

    Decoupled optimisation/filtering visual-inertial odometry fusing RGB frames with inertial measurements for real-time MAV pose estimation.

    Sensors 2023

    Repository

  • B-splines

    Continuous-time trajectory representation with B-spline curves — smooth interpolation for motion control and estimation over asynchronous sensor data.

    Sensors 2023

    Repository

  • HCALIB

    Robust RGB-D–IMU calibration recovering spatial and temporal extrinsics — the alignment step every fusion stack depends on.

    VISAPP 2023

    Repository

Earlier engineering

Aerospace Autonomous Navigation System

M.Sc. project · IBISC, France

Quadrotor stabilisation with collision-avoidance and navigation algorithms in MATLAB and C++.

Flight Simulator — Stewart Platform

B.Sc. graduation project · funded by Boeing

A 6-DOF motion cueing algorithm for pilot training, implemented in C++ and MATLAB/Simulink.

05 / Toolkit

The stack, honestly rated.

Languages

  • Embedded C / C++proficient
  • Pythonproficient
  • MATLAB / Simulinkproficient
  • Rust — scientific & embeddedproficient
  • Javafamiliar

Robotics & middleware

  • ROS 1 / ROS 2experienced
  • PX4 / MAVLinkexperienced
  • eCALexperienced
  • Qt5experienced
  • AUTOSARfamiliar

Estimation & vision

  • OpenCV
  • Ceres
  • GTSAM
  • g2o
  • PCL
  • ORB-SLAM
  • CARLA
  • SciPy · scikit-learn

Deep learning & edge

  • PyTorch
  • TensorRT / CUDA
  • ONNX Runtime
  • NVIDIA Jetson
  • Hailo-8
  • Blaize NPU
  • Raspberry Pi 5

Platforms & tooling

  • Linux
  • Docker
  • Git CI/CD
  • GoogleTest
  • Helix QAC
  • JIRA · Confluence
  • DOORS
  • JTAG · ARM MCU

Standards & safety

  • DO-178B/C
  • ISO 26262
  • ASPICE
  • MLC50/55/60 AUTOSAR

06 / Contact

Let's talk autonomy.

Open to research collaboration, technical advisory and speaking on GNSS-denied navigation.

Full CV and references available on request.