About Me

Machine Learning and Robotics Engineer

Experienced machine learning engineer with demonstrated hands-on experience in industry and academia in machine learning and robotics. Recognised team player in startups and prominent open-source projects, excelling in dynamic, self-directed, and challenging environments.

I am currently working as a machine learning research engineer at Leap AI, Edinburgh, UK.

My current research/work with Leap AI focuses on building multimodal foundation models in embodied AI. The latest application involves using robots for packing.

Objects like fresh produce present a major challenge to current robotic systems: you rarely get two items that look the same (non-rigid, high intra-class variance objects). When packed together in a bin or basket, they are hard to tell apart, and they exhibit a mix of discrete and continuous symmetries, making building a zero-shot foundation model for perception (and manipulation) a fun challenging task.

I have experience with production-ready software and safety-critical systems requiring real-time performance. I have experience working on research projects in industry and taking research work to production. I have experience maintaining software stacks deployed across fleets of active in-field robots. I have an "I can do it then do it" mindset from years of working in startup environment.

I am an open-source advocate. This portfolio is to share my work with everyone. I love applying research in real life. You'll always find me learning about something new.

I am always open to collaboration in research and open-source work. If you find one of my projects interesting or want me on your project, do not hesitate to contact me :)

Résumé & Experience


You can find my career experience on my LinkedIn page and in my résumé available in this section.

Publications & Projects

Click on any project name or image for more details.

Publications

Bayesian Knowledge Tracing For Assessment Results Analysis

This was a collaboration research project between Discovery Education and Uniparticle.
Bayesian Knowledge Tracing, a Hidden Markov model, is utilised to analyse the results of assessments. The technique was tested and validated with a real-life dataset from a coding competition.

Projects

VSLAM Playground

A full Visual SLAM pipeline using state-of-the-art deep learning-based feature descriptors and matchers (SuperPoint, DISK, SuperGlue, LightGlue) and deep stereo depth estimation.

Depth Yolact ROS

A ROS wrapper for yolact instance segmentation with depth image extension for 3D bounding boxes and pointcloud segmentation. The package has different pointcloud filtering techniques (K-means & Gaussian filtering).

IDeepify

A robust deep learning-based Know Your Custmer (KYC) for Egyptian national IDs. (face detection, face matching, Arabic OCR, ID validation)

Deep Computer-Aided Sperm Analysis (CASA)

Faster R-CNN object detection and a modified DeepSORT were used to detect and track human spermatozoa in phase-contrast, dark-field, and bright-field microscopy imaging. The recorded tracks are then used to calculate and report various parameters.

I can only share the given details about this project.

ROS Package: move_base_sequence

A ROS Action server to handle communication with move base action server to navigate through a list of waypoints. The package is available on ROS (Kinetic/Melodic/Noetic) for installation.

D435i VIO

D435i camera is solely used to obtain reliable visual-inertial odometry and SLAM.


Undergraduate Projects

RHex platform

A project made at the robotics lab of University of Central Lanchasire (UCLan). Funded by a grant from Erasmums+.

RowBoat

The project targeted plastic waste in rivers using an autonomous boat for picking the plastic waste from the water.

Automated production line

Undergradaute course project. A fully automated assembly/disassembly low-cost production line. Won the 1st place.

Contact Me

If you have any question, project to work on, or anything to discuss, do not hesitate to contact me on my email or LinkedIn. Click on the email to get the email address copied to your clipboard (click on the envelope in the side panel to get redirected to send an email) or on LinkedIn to get redirected to my LinkedIn account. Thank you in advance.