Charbel Abi Hana — Senior ML Researcher & Engineer

About

I am a Senior AI/ML Engineer at idealworks GmbH in Munich, building the bridge between cutting-edge ML research and production robot systems.

My current work centers on Visual SLAM systems, deploying vision models on edge, managing data infrastructure, leading research projects through Master theses and internships, and using Diffusion Transformers for embodied AI through NVIDIA's Cosmos platform.

I work a transport robot called the iw.hub which is an AMR deployed in over 10 locations with a total of 1200 robots

I hold an M2 in Mathematics, Machine Vision and Machine Learning from École Normale Supérieure Paris-Saclay, one of France's most selective Grandes Écoles.

Experience

Nov 2023 — Present

Senior AI/ML Engineer

idealworks GmbH

Munich, Germany

Mar 2021 — Sep 2023

AI Engineer

inmind.ai

Remote (Company based in Beirut, Lebanon)

Mar 2020 — Mar 2021

Machine Learning Intern

BMW Group / idealworks GmbH

Munich, Germany

Education

2022 — 2023

M2 — Mathematics, Machine Vision & Machine Learning

École Normale Supérieure Paris-Saclay

Gif-sur-Yvette, France GPA: 15.2/20

2021 — 2022

M1 — International Track in Electrical Engineering

Université Paris-Saclay

Évry-Courcouronnes, France GPA: 14.1/20

2015 — 2021

B.Eng — Mechanical Engineering

Holy Spirit University of Kaslik (USEK)

Kaslik, Lebanon GPA: 3.35/4.0

Topics & Focus Areas

NVIDIA Collaboration

NVIDIA Cosmos

Direct collaboration with NVIDIA to fine-tune and integrate NVIDIA Cosmos data generation and augmentation pipeline into the AMR domain and modalities. Developing Synthetic Data Generation pipeline at scale for production use-cases. Tools used include Grafana, Prometheus, LakeFS, Determined AI for GPU inference job orchestration, FastAPI for API development, and Docker for containerization.

predict2.5 transfer2.5 reason2 Fine-tuning Physical AI

NVIDIA Cosmos Webpage→ COMPASS White Paper→

NVIDIA Collaboration

NVIDIA cuVSLAM White paper

Collaborated with the NVIDIA cuVSLAM team to publish a whitepaper on cuVSLAM benchmark and integration in industrial settings where we showcased the performance of cuVSLAM in comparison to other SLAM algorithms. Frameworks used include ROS2, CUDA and NVIDIA Omniverse for Simulation testing.

cuVSLAM VSLAM NVIDIA ROS2 CUDA Omniverse

cuVSLAM→ arXiv→ Webpage→

NVIDIA Collaboration

NVIDIA COMPASS White paper

Collaborated with NVIDIA to publish a whitepaper benchmarking the NVIDIA sim-to-real ecosystem for learned navigation on idealworks robots. Extended the COMPASS benchmark with the iw.hub and a Humanoid Mobile Platform embodiment and a digital twin of the idealworks R&D center, then fine-tuned X-Mobility and COMPASS with NVIDIA Cosmos sim-to-real augmentation and deployed the policies fully onboard the iw.hub's Jetson AGX Xavier, raising real-world success from 65.3% to 91.7%.

COMPASS X-Mobility Cosmos Sim-to-Real Digital Twin Jetson

arXiv→ Webpage→

Production System

Hybrid Visual SLAM

Designed and deployed a hybrid LiDAR + Visual SLAM system for our iw.hub fleet after conducting an elaborate research and investigation into idealwork's LiDAR based localization system shortcomings. The system is currently in pilot mode and is expected to be deployed and scaled to full on large-scale production system by end-of-year. Designed a benchmarking framework deployed on-site and on AWS for system comparison and agile updates. Frameworks used include ROS2, CUDA, FastAPI, AWS EC2, AWS Lambda, AWS Batch, AWS ECR,Github Actions, Bazel.

VSLAM LiDAR ROS2 CUDA FastAPI AWS Bazel

Production System

Anonymization Pipeline

Built and deployed a modular people anonymization pipeline for autonomous mobile robots in production. The pipeline support modular model support with automatic benchmarking and selection. Deployed on-site and on AWS with Lambda, EC2, Batch and S3. CI/CD with Github Actions. On-site deployment uses Docker and Celery for distributed processing.

Anonymization Segmentation AWS

Production System

Motion Prediction for Collision Avoidance

Built and deployed a robot motion prediction pipeline on-edge on NVIDIA Jetson Xavier device for on-edge collision avoidance in production. Technologies for the core were a YOLO model for object detection and an Unscented Kalman Filter for motion estimation and tracking. Deployment was done using ROS, Docker, ONNX and TensorRT.

Collision Avoidance Motion Prediction On-Edge Deployment ROS ONNX TensorRT

Production System

Data Infrastructure

Built infrastructure and pipelines for robot data collection and processing in production which includes people anonymization, ROS2 bags handling, version control and data storage. Building large industrial datasets for training world models and other perception models. Built an extension on top of Rerun for custom data visualization, analysis and annotation. Initiated the Data Driven Decision Making (DDDM) framework for better problem understanding and solutions delivery. Frameworks used include ROS2, DVC and LakeFS for data version control, NAS and AWS S3 for data storage.

DVC LakeFS NAS AWS S3

Publications

NVIDIA White paper

"Physical AI for Intralogistics: Benchmarking the NVIDIA Sim-to-Real Ecosystem for Cosmos-Enhanced COMPASS Navigation on Industrial Mobile Robots"

Charbel Abi Hana, Rayan Zaitouni, Kameel Amareen, Mohamad Mostafa, Hasan Sayour, Wei Liu, Yan Chang, Mihir Acharya, Anthony Rizk

Collaboration with NVIDIA · First Author

Benchmarks NVIDIA's X-Mobility and COMPASS workflow for learned navigation on Idealworks robots. The COMPASS benchmark is extended with the iw.hub and a Humanoid Mobile Platform embodiment and a digital twin of the Idealworks R&D center; policies trained in the digital twin, with and without NVIDIA Cosmos sim-to-real augmentation, are deployed on the iw.hub's onboard Jetson AGX Xavier and compared against the production navigation stack. Cosmos augmentation raises real-world success from 65.3% to 91.7% and nearly halves weighted travel time, bringing learned navigation within 8.3 points of the production stack, fully onboard.

arXiv · Webpage

NVIDIA White paper

"Industrial cuVSLAM Benchmark & Integration"

Charbel Abi Hana, Kameel Amareen, Mohamad Mostafa, Dmitry Slepichev, Hesam Rabeti, Zheng Wang, Mihir Acharya, Anthony Rizk

Collaboration with NVIDIA · First Author

A comprehensive benchmark of visual odometry and Visual SLAM systems for mobile robot navigation in real-world logistics environments. Multiple VO approaches are evaluated across controlled translational, rotational, and mixed-motion trajectories, as well as a large-scale ~1.7 km production facility dataset, using Absolute Pose Error against Vicon motion-capture and LiDAR-SLAM ground truth. Results show that a hybrid stack combining the cuVSLAM front-end with a custom SLAM back-end achieves the strongest mapping accuracy, motivating its deeper integration as the core VO component. The stack is further validated through deployment on an NVIDIA Jetson platform.

arXiv · Webpage · cuVSLAM Github · NVIDIA Press Release

AAMAS 2026

"SPADE: Sketch-guided Path Planning Augmented with Diffusion Experts"

Charbel Abi Hana, Tatiana Ghantous, Mikael Khalil, Anthony Rizk

Extended Abstract · First Author

Addresses key limitations of imitation-learning-based path planning for AMRs—limited generalization to unseen environments and fragile demonstration collection—through two main contributions: an overhauled ROS 2 annotation tool and a novel training strategy that integrates diffusion-based augmentation into baseline behavioral cloning models. Evaluated via ablation studies on a provided expert-demonstration dataset, the enhanced approach outperforms state-of-the-art methods, achieving 39.1% lower Absolute Pose Error (APE) and 33.5% lower Fréchet Inception Distance (FID) while requiring 93.8% fewer trainable parameters, and attains diffusion-level generalization while preserving real-time, on-edge inference properties.

arXiv · OpenReview · Webpage · Code (forthcoming)

IPAS 2025

"End-to-end Sketch-Guided Path Planning through Imitation Learning for Autonomous Mobile Robots"

Anthony Rizk, Charbel Abi Hana, Youssef Bakouny, Flavia Khatounian

IEEE IPAS · Second Author

Introduces a more accessible and flexible approach to AMR path planning through sketch-guided imitation learning, where non-technical users simply draw the desired navigational path on a provided 2D map to teach U-net models path planning behaviors—replacing traditional reward-based learning or costly hardware-dependent techniques. A novel evaluation framework combining metrics from image generation and robotics is proposed, and the approach is integrated into an end-to-end robotics stack to demonstrate practical usability.

IEEE Xplore · Webpage · Code (SKIPP)

Open Source Projects

Open-source work and research implementations. View all repositories on GitHub.

Research & Robotics

SKIPP

End-to-end Sketch-Guided Path Planning through Imitation Learning for Autonomous Mobile Robots

Python View Repository

Implementation of the publication "End-to-end Sketch-Guided Path Planning through Imitation Learning for Autonomous Mobile Robots".

Imitation LearningPath PlanningAMRPyTorchDockerOccupancy Grids

SidewalkDetection

Semantic segmentation for autonomous robot navigation in urban environments

Python View Repository

Pixel-level semantic segmentation enabling robots to identify safe sidewalk areas from RGB camera feeds. Complete pipeline from data labeling through training to real-time inference with a custom 276-image dataset.

Semantic SegmentationU-NetComputer VisionOpenCVNavigation

ups-marl-benchmark

Multi-Agent Reinforcement Learning benchmarking for decentralized autonomous driving

Python View Repository

Benchmarking suite for evaluating decentralized multi-agent RL algorithms in high-density autonomous driving scenarios using highway-env simulation.

Multi-Agent RLDecentralized Controlhighway-envAutonomous Driving

Machine Learning

ups-mv-gans

GANs for data augmentation and image-to-image translation

Python View Repository

Tackling data scarcity through GAN-based synthetic image generation and CycleGAN for unpaired image-to-image translation across visual domains.

GANsCycleGANData AugmentationGenerative AIDocker

ups-ml-football-player-value

Regression analysis for predicting professional athlete market values

Python View Repository

End-to-end ML project predicting football player transfer values. Demonstrates reproducible ML workflows with DVC for experiment tracking and Dockerized execution.

RegressionDVCDockerScikit-learnMLOps

ups-ml-stroke-prediction

Healthcare analytics: classification model for stroke risk prediction

Python View Repository

Classification project for predicting stroke risk using patient health data. Covers EDA, class imbalance handling, and multi-classifier comparison.

ClassificationHealthcareEDAScikit-learnData Visualization

Tools & Infrastructure

gpu-test

Utility scripts for verifying GPU acceleration across TensorFlow 2 and PyTorch environments.

View on GitHub

astronvim-setup

Customized Neovim configuration built on AstroNvim, optimized for Python, C++, and CUDA development.

View on GitHub

dvc-basics

Data and model version control using DVC in MLOps — setup guides, remote storage, and Python API integration.

View on GitHub

tex-templates

Professional LaTeX templates for academic publications and conference papers (IEEE format presets).

View on GitHub

Technical Skills

AI & Robotics

Diffusion Policy VLA Imitation Learning Computer Vision VSLAM Motion Prediction Edge Deployment

Stack

Python C++ CUDA PyTorch NVIDIA Cosmos NVIDIA Isaac ROS2 Docker

Infrastructure & Languages

GPU Clusters Distributed Training Linux Kubernetes

Get in Touch

I'm always happy to chat about robotics, diffusion models, or anything AI. Feel free to reach out, whether it's a collaboration, a question, or just to say hi 👋

Email GitHub LinkedIn