I'm a Machine Learning Engineer with a strong foundation in Computer Vision and Deep Learning, holding an MSc in Informatics from the Technical University of Munich (TUM). I'm passionate about developing and deploying intelligent systems to solve real-world challenges.
- 🎓 MSc Informatics @ Technical University of Munich (TUM) (GPA 1.5)
- Thesis: FollowMe: Human Detection and Tracking in The Field based on PointPillars with Range Image Feature Fusion and Unsupervised Domain Adaptation (Grade 1.3)
- 🎓 BSc Computer Science @ The German University in Cairo (GUC) (GPA 1.4)
- Thesis: Photorealistic Rendering of Training Data for Object Detection and Pose Estimation with a Physics Engine (Grade 1.0)
- 🌱 Currently working as an AI NLP Intern at CONXAI, contributing to document analysis frameworks using LLMs and RAG.
- 💡 Previously worked on Computer Vision tasks (Object/Material ID with YOLO/SAM) at CONXAI, managing the ML lifecycle from data to deployment.
- 🚀 Experienced in 3D data processing (LiDAR, Depth), feature fusion, domain adaptation, and deploying ML models in various contexts (robotics, autonomous systems).
- Machine Learning / Deep Learning: Computer Vision, Deep Learning, Machine Learning, RAG, Object Detection/Tracking, Pose Estimation, 3D Reconstruction
- Frameworks / Libraries: PyTorch, LangChain, OpenCV, BlenderProc
- Programming: Python, C++
- Tools & Platforms: AWS, ROS2, Docker, Django, FastAPI
- Languages: English (C1), German (B1), Arabic (Native)
- AI NLP Intern @ CONXAI (Dec 2024 - Present): Document analysis using LLMs & RAG.
- AI Computer Vision Intern @ CONXAI (Mar 2024 - Nov 2024): Object/material identification (YOLO, SegmentAnything), ML lifecycle management.
- Master Thesis Student @ Fraunhofer IPA (Jul 2023 - Jan 2024): Human detection/tracking (LiDAR), feature fusion, unsupervised domain adaptation (achieved +7% in mAP improvement).
- R&D Working Student @ Blickfeld (Oct 2022 - Mar 2023): Point cloud noise filtering evaluation, ROS2 test framework development.
- IDP Student - Computer Vision @ Angsa Robotics (Apr 2022 - Oct 2022): Real-time obstacle detection algorithm development from depth images.
- Software Developer @ Nebumind GmbH (May 2021 - Sep 2022): Contributed to a custom spatiotemporal database.
- Guided Research (TUM): Category-Level 6d Pose Estimation Using Monocular Depth Estimation - Network and loss design experimentation.
- Advanced Topics in 3D Computer Vision (TUM Lab): RGB-based 6D Pose Estimation using inferred depth (GPVPose, DPTMonoDepth) and synthetic data (BlenderProc).
- Advanced Deep Learning for Computer Vision (TUM Course): Weakly-supervised 3D mesh reconstruction from sparse point clouds.
- Bachelor Thesis (AT TUM CAMPAR): Synthesizing photorealistic training data (Arnold, Nvidia Physx) for 6D pose estimation, exploring domain adaptation.
(Feel free to link specific project repositories here if they are public!)
- Email: wessam.frrag@hotmail.com
- LinkedIn: linkedin.com/in/wessam-frrag