Home

Curriculum Vitae

Education, experience, publications, and awards.

Summary

Ph.D. candidate at the Institute for Visual & Analytical Computing specializing in Multimodal Perception & Robotics. Proven track record of building real-world robotic systems for logistics and industrial bin-picking, leveraging synthetic data generation and monocular depth estimation pipelines. Passionate about advancing research in Embodied AI and trustworthy Human-Robot Interaction (HRI) through the convergence of Spatial Intelligence and Mixed Reality interfaces.

Education

2025 –
Ph.D., Computer Science — Universität RostockRostock, Germany
2020–25
M.Sc. Robotics — Technische Universität Dortmund GPA 1.90Dortmund, Germany
2022
Exchange Semester, M.Sc. AI & Robotics — Sapienza Università di RomaFeb – Jul 2022 · Rome, Italy
2016–20
B.Sc. Electrical Engineering & Computer Science — Czech Technical University in Prague GPA 1.60Prague, Czech Republic

Experience

Research Associate, Institute for Visual & Analytical Computing (VAC) — Universität Rostock

Rostock, Germany
  • Leading research in multimodal LLMs, egocentric perception for robotics, and computer vision as part of CORE Labs.
  • Collaborating with industry partners (Ramblr & NeoBIM) to develop vision-language models (VLMs) that serve as intuitive interfaces for complex software and robotic systems.
  • Supervising student research groups and mentoring research assistants (HiWis) in deep learning and computer vision methodologies.

Research Assistant in Robotics & AI — Fraunhofer IML

Dortmund, Germany
  • Master's Thesis: developed a perception method for real-time volume feedback in partnership with DB Schenker, directly enhancing logistics throughput.
  • Generated synthetic datasets (NVIDIA Isaac Sim) and curated real-world RGB-D data to fine-tune ViT and ConvNeXt depth-estimation networks.
  • Engineered a parallelized RGB-D processing pipeline (52% speedup) and deployed ROS nodes for calibrated LiDAR/stereo-camera integration.

Working Student, Computer Vision & Machine Learning — Swisslog Robogistics Technology Center

Dortmund, Germany
  • Developed an auto-annotation tool leveraging Meta's Segment Anything Model (SAM) to curate 600+ RGB-D scenes for the ItemPiQ bin-picking system.
  • Trained a Bag-of-Visual-Words material classifier achieving 70% accuracy on unseen parts with sub-7ms inference time.
  • Designed vision-based suction grasping algorithms for cluttered bins, benchmarking cameras and DL models to optimize performance.

Working Student, Bin-Picking — Lehrstuhl für Fördertechnik (FLW), TU Dortmund / Fraunhofer IML

Dortmund, Germany
  • Curated the DoPose dataset (COCO format) by engineering an ICP-stitching pipeline for Zivid 2 point clouds, and benchmarked Detectron2 for segmentation accuracy.
  • Prototyped low-level grasping routines (C++/Python) and deployed trajectory planning via KUKA MoveIt! on a two-finger gripper.

Undergraduate Researcher in Robotics — Multi-Robot Systems Group, Czech Technical University

Prague, Czech Republic
  • Bachelor's Thesis: designed a novel UAV grasping mechanism and led flight-testing, securing 1st place at the MBZIRC 2020 Challenge II.
  • Engineered a closed-loop grasping system using ROS and embedded C/C++ (ATMega328) for real-time sensor fusion and aerial manipulation.

Publications

  1. 2026

    MCP4IFC: IFC-based Building Design using Large Language Models

    B. K. Nithyanantham, T. Sesterhenn, A. Nedungadi, S. Peral Garijo, J. Zenkner, C. Bartelt, S. Lüdtke

    EC3 — European Conference on Computing in Construction, Corfu, Greece accepted

  2. 2026

    RoboBIM: Scaling Robot Semantic Foundations through Agentic Extraction of BIM Priors

    A. Nedungadi, et al.

    ICRA 2026 Workshop on Robots Meet Prior Maps, Vienna, Austria workshop

  3. 2026

    Benchmarking 3D Reconstruction for Under-Ice Robotic Perception in Arctic Environments

    A. Nedungadi, et al.

    ICRA 2026 Workshop on Sea to Space, Vienna, Austria workshop

  4. 2026

    BIM-Edit: Benchmarking Large Language Models for IFC-Based Building Information Modeling

    B. K. Nithyanantham, C. Kujat, T. Sesterhenn, S. Telgmann, A. Nedungadi, J. Plönnigs, C. Bartelt, S. Lüdtke

    arXiv preprint arXiv:2606.20146 · under review for NeurIPS 2027 under review

  5. 2025

    Monocular Depth Estimation for Efficient Volume Reconstruction in Logistics

    A. Nedungadi, C. Rest, J. Stenzel

    Manuscript in preparation in prep

  6. 2023

    3D Models of the Martian Surface in Virtual Reality

    A. Nedungadi, C. Wöhler, et al.

    AAS / Division for Planetary Sciences Meeting Abstracts, 55, 212.04 conference

  7. 2022

    DoPose: Dataset for Object Segmentation and 6D Pose Estimation

    A. Gouda, A. Nedungadi, et al.

    Zenodo · DOI 10.5281/zenodo.6103779 dataset

  8. 2020

    Design of an Active-Reliable Grasping Mechanism for Autonomous Unmanned Aerial Vehicles

    A. S. Nedungadi, M. Saska

    Modelling and Simulation for Autonomous Systems (MESAS), Springer LNCS, pp. 162–179 conference

Full list on Google Scholar →

Skills

LanguagesPython · C++ · C · Rust · MATLAB
Robotics & PerceptionROS2 · MoveIt! · NVIDIA Isaac Sim · SLAM · 3D Reconstruction · Monocular Depth Estimation · PCL · Zivid SDK
Libraries & FrameworksPyTorch · Hugging Face Transformers · MCP · Detectron2 · OpenCV · Open3D · scikit-learn · NumPy · Pandas · SciPy
DevOps & ToolsDocker · Git · CI/CD · CMake · Linux · Jupyter · Pinecone · MongoDB

Honors & Awards

Leadership & Extracurricular

Participant — Digital Entrepreneurship Lab (2025 Cohort), Technische Universität Dortmund

  • Selected for a competitive innovation program to co-found a logistics venture utilizing 3D perception and mixed reality.
  • Winner of the final pitch competition for optimizing 1PL–3PL workflows with the "Packzilla" startup concept.

Vice President — Erasmus Student Network Dortmund

  • Directed a team of 50+ volunteers to execute large-scale cultural events, enhancing international student integration.
  • Steered board-level strategy for PR, partnerships, and recruitment; managed budget allocation for semester programs.

President — Fachschaft Automation & Robotics, Technische Universität Dortmund

  • Elected representative for M.Sc. students, acting as the primary liaison to faculty and industry partners.
  • Spearheaded onboarding for 60+ incoming students and initiated a monthly newsletter to drive community engagement.

Languages, Interests & Affiliations

Languages: English (C1), German (B1), Spanish (A2), Italian (A1). IEEE Member.

Interests: hiking, competitive poker, SCUBA diving, reading, painting.