Research Software Engineer | AI & Robotics Enthusiast | Gamer

Hi, I'm Saifullah, a research software developer and PhD candidate at the German Research Center for Artificial Intelligence (DFKI) GmbH, Kaiserslautern, Germany. My current research work interests include Explainable AI, Privacy-Preserving Machine Learning, and Model Robustness in Document Intelligence. Outside of research, I enjoy robotics and gaming.

  • Google Scholar icon
    • Download Resume
    Work
    1. German Research Centre for Artificial Intelligence (DFKI) GmbH
      PhD Researcher
      —
    2. Intelligent Robotics Lab (IRL), National Center of Artificial Intelligence (NCAI)
      Technical Lead / Research Associate
      —
    3. Ingenio
      Lead Robotics
      —
    4. Robotics & Intelligent Systems Engineering (RISE) Lab, NUST
      Team Lead Project RoboCup
      —
    5. Robotics & Intelligent Systems Engineering (RISE) Lab, NUST
      Robotics Software Intern
      —
    6. NUST Aerial Robotics (NUSTAR) Lab, NUST
      Junior Researcher
      —
    Education
    1. RPTU University of Kaiserslautern-Landau
      Doctorate (PhD) in Computer Science (CS)
      —
    2. National University of Sciences and Technology (NUST)
      MS Robotics & Intelligent Machine Engineering (RIME) · GPA: 1.0
      —
    3. National University of Sciences and Technology (NUST)
      Bachelors (BS) in Mechanical Engineering · GPA: 1.31
      —
    Exchange Programs & Competitions
    1. Oxford Machine Learning Summer School (OxML) 2023
      Summer School Participant
      —
    2. Sakura Exchange Program in Science, Tohoku University, Japan
      Research Exchange Student
      —
    3. Standard Platform League (SPL), RoboCup 2016
      Team Member, Team NUST
      —
    Skills
    Deep Learning
    PyTorchPyTorch IgniteExplainable AIPrivate MLDocument AICNNsTransformersDiffusion Models
    Frontend/Backend
    FastApiReactNextJSS3LakeFSSQL-AlchemySupabaseRabbitMQPySparkDockerGit
    LLM Frameworks
    LangChainRAGvLLM
    Database
    PostgreSQLMongoDB
    Programming
    PythonC++CJavascript
    Operating Systems
    LinuxDockerWindowsGentoo
    Robotics
    ROSGazeboVREPOpenCVMulti-body DynamicsMotion PlanningSLAM
    Edge AI
    NVIDIA DeepstreamNVIDIA JetsonTensorRT
    Graphics Designs
    Adobe Photoshop
    Publications
    1. Are Deep Models Robust against Real Distortions? A Case Study on Document Image Classification.
      Saifullah, S., Siddiqui, S. A., Agne, S., Dengel, A., Ahmed, S.
      26th International Conference on Pattern Recognition (ICPR), 2022, IEEE
    2. Analyzing the Potential of Active Learning for Document Image Classification.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      International Journal on Document Analysis and Recognition (IJDAR), vol. 26, no. 3, pp. 187--209, 2023. Springer.
    3. DocXClassifier: Towards a Robust and Interpretable Deep Neural Network for Document Image Classification.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      International Journal on Document Analysis and Recognition (IJDAR), pp. 1--27, 2024. Springer.
    4. ColDBin: Cold Diffusion for Document Image Binarization.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      International Conference on Document Analysis and Recognition, pp. 207--226, 2023. Springer Nature Switzerland, Cham.
    5. PrIeD-KIE: Towards Privacy Preserved Document Key Information Extraction.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      arXiv preprint arXiv:2310.03777, 2023.
    6. The Privacy–Explainability Trade-off: Unraveling the Impacts of Differential Privacy and Federated Learning on Attribution Methods.
      Saifullah, S., Mercier, D., Lucieri, A., Dengel, A., Ahmed, S.
      Frontiers in Artificial Intelligence, vol. 7, 1236947, 2024. Frontiers Media SA.
    7. Towards Privacy Preserved Document Image Classification: A Comprehensive Benchmark.
      Saifullah, S., Mercier, D., Agne, S., Dengel, A., Ahmed, S.
      International Journal on Document Analysis and Recognition (IJDAR), pp. 1--25, 2024. Springer.
    8. The Reality of High Performing Deep Learning Models: A Case Study on Document Image Classification.
      Agne, S., Dengel, A., Ahmed, S.
      IEEE Access, 2024. IEEE.
    9. DocXplain: A Novel Model-Agnostic Explainability Method for Document Image Classification.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      arXiv preprint arXiv:2407.03830, 2024.
    10. StylusAI: Stylistic Adaptation for Robust German Handwritten Text Generation.
      Riaz, N., Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      International Conference on Document Analysis and Recognition, pp. 429--444, 2024. Springer Nature Switzerland, Cham.
    11. Latent Diffusion for Guided Document Table Generation.
      Hamdani, S. J. H., Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      International Conference on Document Analysis and Recognition, pp. 368--383, 2024. Springer Nature Switzerland, Cham.
    12. DocVCE: Diffusion-based Visual Counterfactual Explanations for Document Image Classification.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      arXiv preprint arXiv:2508.04233, 2025.
    13. DP-DocLDM: Differentially Private Document Image Generation using Latent Diffusion Models.
      Saifullah, S., Agne, S., Dengel, A., Ahmed, S.
      arXiv preprint arXiv:2508.04208, 2025.