Curriculum Vitae
Amin Dariani
Research focus
Explainable and trustworthy machine learning — making what a model has learned inspectable, and checkable rather than simply believed. Background in computer vision and eye tracking, robot perception and model-based control.
IEducation
M.Sc. Automation and Control
Current average 2.1. Specialisation in Connected Automation Systems.
Thesis (in progress) — explaining graph neural networks. Advisor: M. Becker, Fraunhofer IOSB / KIT IES.
Selected modules
- Optimal Controlgrade 1.3
- Fault Diagnosis and Fault-Tolerant Controlgrade 1.3
- Modelling and Identificationgrade 1.7
- Cooperative Robot Controlgrade 2.0
- Master Project CASgrade 1.3Recognising industrial activity from gaze →
- CAS Project LabVision-guided pick-and-place with a Franka Emika Panda →
B.Sc. Mechanical Engineering
Overall grade 2.3.
Thesis — “Under-actuated Double Inverted Pendulum Control using LQR, PID and Fuzzy Control”.
Selected modules
- General Mathematics Igrade 1.0
- Robotics, with laboratorygrade 2.0
- Numerical Computationgrade 2.0
- Automatic Controlgrade 2.3
IIResearch experience
Working Student Researcher, Explainable AI
Implementing explainable-AI methods inside analysis tooling, making model decisions inspectable by non-ML users.
Student Research Assistant, Gaze‑based Activity Recognition
Gaze-based human activity recognition for industrial settings: study design, data collection on HoloLens 2, and evaluation of classification models.
Undergraduate Research Assistant
Adaptive tracking control on a generalised fuzzy hyperbolic model for FUME, a lower-limb exoskeleton for paraplegic users; validated on the physical robot against a tuned PID baseline. First-author paper at ICRoM 2019.
IIIPublications
- Nov 2019
A. Amir-B.D., S. M. Tahamipour-Z., A. Akbarzadeh. “Adaptive Tracking Control Based on GFHM for a Reconfigurable Lower Limb Exoskeleton.” 7th International Conference on Robotics and Mechatronics (ICRoM), Tehran, 20–21 November 2019, pp. 74–79.
- In preparation
A. Amir-B.D., S. Walunj. Manuscript on gaze-based classification of industrial activities from HoloLens 2 eye tracking, 2026.
IVSelected projects
- Mar 2026
Eye Movement Classification from HoloLens 2 Eye Tracking
Designed and ran the data collection in the DFKI smart factory and compared classifiers for recognising industrial activities from gaze.
- Winter 2024/25
Vision-guided Pick-and-Place with a Franka Emika Panda
ROS system in which a 7-DOF Franka Emika Panda locates a randomly placed object, plans a collision-free path around static obstacles and transfers it to a target pose. Trained the object detector on a dataset I recorded and annotated.
VTechnical skills
Programming
- Python
- MATLAB/Simulink
- JavaScript
Machine learning and computer vision
- PyTorch
- scikit-learn
- OpenCV
- YOLO
Robotics and AR
- ROS
- MoveIt
- Franka Emika Panda
- Intel RealSense
- HoloLens 2
- Unity
- MRTK
- ARETT
Methods and models
- Graph neural networks
- Optimal and model-predictive control
- Adaptive control
- System identification
Languages
- English — fluent (IELTS 8.0)
- German — B1
- Persian — native
VIOther professional experience
- Jan 2024 – Jan 2026
CAP Developer, Working Student
- Aug 2021 – Aug 2023
Senior Front-End Developer
- Sep 2019 – Aug 2021
Front-End Developer
VIIReferences
“The fact that he carried out this project from the initial idea through implementation and toward an IEEE publication demonstrates his independence, persistence, and strong potential as a researcher.”
“He is now a key member of the FUM Robotics Research Lab and has been an active participant in various teams and has definitely been an asset to our laboratory.”
Further references, from my master’s project advisor at RPTU and my thesis supervisor at Fraunhofer IOSB, are available on request.