Farhad Hoseyni

I'm

About

AI and Robotics Engineer working on computer vision, robot learning, and practical intelligent systems.

I work on computer vision, deep learning, and robotic manipulation to solve real-world problems in industrial automation and healthcare. Currently, I am pursuing my Master's in AI and Robotics at the University of Twente and working as a Computer Vision and Robotic Engineer at the Fraunhofer Innovation Platform.

My background combines classical control engineering (B.Sc. in Electrical and Control Engineering from K. N. Toosi University of Technology with a minor in Computer Engineering) with practical machine learning: building Vision-Language-Action (VLA) robot policies, sub-millimeter 3D bin-picking systems, and clinical triage tools for medical imaging.

Download My CV

Education

K. N. Toosi University of Technology

Bachelor's Degree in Electrical and Control Engineering
Minor Degree in Computer Engineering
Sep. 2019 to Sep. 2024
Grade: 18.33 / 20 (Ranked 6th in entrance)

University of Twente

Master's Degree in AI and Robotics
Specialization in Software and Algorithm AI
Sep. 2025 to Present
Enschede, Netherlands

Professional Experience

Industry roles and technical projects across robotics, computer vision, and medical imaging

Fraunhofer Innovation Platform

CV & Robotics Engineer

Oct. 2025 - Present • Enschede, NL

Smartory Labs

AI & Computer Vision Engineer

Mar. 2023 - Sep. 2024 • Tehran, IR

APAC Research Group

Technical Manager & CV Engineer

Jun. 2022 - Sep. 2025 • Medical AI

Fraunhofer Innovation Platform

Computer Vision and Robotic Engineer

Oct. 2025 to Present

Enschede, Netherlands

Built and deployed a multi-modal battery inspection system combining optical surface cameras, electrochemical tests, and a Yaskawa handling arm, reaching 95% defect recall and cutting inspection cycle times by 50%.
Programmed Universal Robots (UR) and Yaskawa arms for modular assembly operations including pick-and-place, precision screwdriving, welding, and in-line quality inspection.
Integrated Zivid 3D depth cameras and Pick-it 3D vision processors into the robot loop, achieving <0.8 mm grasping precision despite part shifts and heavy industrial glare.
Conducting ongoing research on Vision-Language-Action (VLA) controllers to let robots execute assembly sequences directly from operator text instructions and camera feedback.

Key Initiatives

Battery Defect AI Inspection 95% Recall Flexible UR & Yaskawa Assembly Zivid & Pick-it 3D Manipulation <0.8mm Precision VLA Controller Research (Ongoing)

View Fraunhofer Project Page

Smartory Labs

AI and Computer Vision Engineer

Mar. 2023 to Sep. 2024

Istanbul, Turkey

Directed end-to-end clinical data operations, coordinating the collection and multi-stage annotation of 100,000+ dental radiographs with a team of 20+ dental specialists.
Developed an end-to-end diagnostic assistant for dental radiography, detecting 60% of hidden interproximal cavities and restorations that can be missed during routine checks.
Built automated patient profiling tools that update clinical charts and track oral hygiene over time using everyday smartphone photos.
Created and open-sourced AugmenTory, a Python polygon augmentation library that reduces spatial memory usage to 1.2% during instance segmentation training.

Key Initiatives

100K+ Radiographs Dataset 20+ Specialists AI Diagnostic Assistant 60% Anomaly Catch Daily Hygiene & Profiling AugmenTory Framework Open-Source

View Smartory Labs Project Page

APAC Research Group

Technical Manager and Computer Vision Engineer

Jun. 2022 to Sep. 2025

Medical AI Research

Developed deep learning triage pipelines for intracranial hemorrhage diagnosis on non-contrast head CT scans and tumor detection on brain MRI.
Leveraged generative AI and self-supervised foundation representations to dramatically reduce reliance on manual pixel-level annotations, combining them with Grad-CAM heatmaps to deliver interpretable, clinical-grade decision support.
Curated a 300-patient CT scan research dataset (Hemorrica) and developed a web-based medical image annotation tool for multi-specialist labeling.