Viet Pham

About

My research sits at the intersection of medical robotics, computer vision, and machine learning — spanning reinforcement and imitation learning for surgical task automation, vision-language(-action) models for robotics, and applied computer vision in surgical scenes.

My path runs from data engineering at apic.ai, through surgical-robotics research at KIT's HERA Lab (under Prof. Franziska Mathis-Ullrich) and Carl Zeiss AG, to a visiting scholarship at Johns Hopkins (IMERSE Lab, Prof. Axel Krieger), where I built multimodal imitation-learning models for robotic partial nephrectomy.

Currently, I'm completing my M.Sc. in Computer Science at TUM, with a focus on AI in Medicine and Computer Aided Medical Procedures.

Research Interests

• Surgical robotics — autonomous and semi-autonomous control for minimally invasive procedures
• Imitation & reinforcement learning — learning manipulation policies from demonstration and simulation
• Vision-language(-action) models — grounding natural-language instructions in robotic actions
• Computer vision for surgery — instrument segmentation, scene reconstruction, intra-operative perception

Publications

Authors in bold indicate myself. See also my Google Scholar profile.

1. Open-H Embodiment: The first healthcare robotics open dataset

   Open-H Embodiment Consortium (Dataset Contributor)

   NVIDIA, Johns Hopkins University, Technical University of Munich · April 2026

   arXiv:2604.21017

2. Domain-agnostic weakly supervised surgical instrument segmentation

   Rebekka Charlotte Peter, Doan Xuan Viet Pham, et al., Franziska Mathis-Ullrich

   Scientific Reports (Nature Portfolio) · March 2026

   doi:10.1038/s41598-026-43054-1

3. Highly Accurate Real-time Surgical Scene Rendering using Gaussian Surfels

   Idris O. Sunmola, et al., Doan Xuan Viet Pham, Axel Krieger

   IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2026) · November 2025

   ★ Best Paper Award · SPAR3D Workshop

   arXiv:2503.04079

4. Lens Capsule Tearing in Cataract Surgery Using Reinforcement Learning

   R. C. Peter, S. Peikert, L. Haide, D. X. V. Pham, et al., F. Mathis-Ullrich

   IEEE International Conference on Robotics and Automation (ICRA 2024) · August 2024

   doi:10.1109/ICRA57147.2024.10611714

Experience

1. Apr 2025 — Sep 2025

   Visiting Graduate Scholar

   Johns Hopkins University · IMERSE Lab (Prof. Axel Krieger) · Baltimore, USA

   • Developed multimodal imitation learning models for robotic partial nephrectomy, fusing visual observations with BERT-encoded language instructions and pre-operative anatomical context for action prediction.
   • Contributed multimodal surgical demonstration data (robot trajectories, synchronized camera streams) to the Open-H Embodiment dataset — a multi-institution open-source healthcare robotics initiative (NVIDIA, JHU, TUM).
2. Oct 2023 — Mar 2025

   Working Student · Research & Innovation

   Carl Zeiss AG · Karlsruhe, Germany

   • Self-initiated interdisciplinary research collaboration with the AI in Medicine (AIM) Lab at TUM on instrument localization in microscope-integrated OCT images; implemented methodology and conducted experiments underlying the Scientific Reports 2026 publication.
   • Extended and maintained reinforcement learning pipelines in medical simulations (SOFA); integrated new features and conducted follow-up experiments.
3. Jun 2023 — Sep 2023

   Research & Innovation Intern

   Carl Zeiss AG · Karlsruhe, Germany

   • Designed and trained reinforcement learning agents for the capsulorhexis procedure in cataract surgery; implemented the SOFA-based RL training pipeline, designed the reward function, and conducted the experiments underlying the ICRA 2024 publication.
   • Ran wet-lab experiments on a Mecademic single-arm robot using ex-vivo porcine tissue and phantoms, including software/hardware integration for closed-loop agent evaluation.
4. Jun 2022 — Mar 2023

   Student Research Assistant

   Karlsruhe Institute of Technology · HERA Lab · Karlsruhe, Germany

   • Performed semantic image segmentation on surgical images using deep learning.
   • Ported ETH Zurich's magnetic continuum robot (m-CR) SOFA simulation for reinforcement learning.
5. Jan 2020 — Mar 2022

   Working Student · Software / Data Engineer

   apic.ai GmbH · Karlsruhe, Germany

   • Developed ETL data pipelines for AI-based video analysis, processing over 35 TB / year.
   • Optimized Kubernetes clusters with up to 256 NVIDIA GPU instances (GKE, EKS).

Education

1. Oct 2023 — Aug 2026

   M.Sc. Computer Science

   Technical University of Munich · Munich, Germany

   Focus: AI in Medicine, Computer Aided Medical Procedures, Computer Vision.

2. Apr 2025 — Sep 2025

   Master's Thesis in Computer Science

   Johns Hopkins University · Baltimore, USA

   Multi-Modal Imitation Learning for Robotic Partial Nephrectomy · Grade: 1.0

3. Oct 2017 — Sep 2023

   B.Sc. Computer Science (Minor: Electrical Engineering)

   Karlsruhe Institute of Technology · Karlsruhe, Germany

   Thesis: Reinforcement Learning for Laparoscopic Tissue Retraction · Grade: 1.0

Skills

Tooling

Python · C++ · C# · PyTorch · ROS · OpenCV · NumPy · SOFA · Docker · Kubernetes

Methods

Reinforcement Learning · Imitation Learning · Vision-Language Models · Computer Vision · Medical Robotics

Languages

German (native) · Vietnamese (native) · English (fluent)