Principal Research Developer · Computer Vision

I build intelligent systems that move from research to the real world.

Principal research developer at Autodesk with 15+ years across AI, computer vision, VLMs, perception, sensor fusion, and production software—turning ambitious ideas into scalable systems.

Portrait of Mahdi Marsousi
Toronto, Canada
15+Years building AI systems
1M+Labeled images in major programs
1,000+Deployed enterprise sites
8Selected patent families

The throughline

From volumetric medical imaging to city-scale monitoring and multimodal perception, my work sits where scientific rigor, software architecture, and operational reality meet.

Selected work / 01—07

Systems built for complexity.

Representative platforms spanning health, mobility, safety, and intelligent infrastructure.

04ML Infrastructure

AI Computer Vision Platform

A modular workflow platform for composing, training, evaluating, and deploying computer-vision pipelines with C++ performance and Python extensibility.

  • Workflow engine
  • C++ / Python
  • Edge + cloud
05Sensor Fusion

Automated Parking System

Vehicle-environment perception using ultrasonic sensors and cameras for parking-space detection, self-alignment, collision avoidance, and automated maneuvering.

  • Camera + ultrasonic
  • Geometry
  • State estimation
06Safety & Operations

Smart Swimming Pool

RGB and infrared perception for swimmer detection, segmentation, and tracking—supporting safety monitoring and intelligent fresh-water operations across four countries.

  • RGB + infrared
  • Segmentation
  • Human tracking
07Medical Imaging

Volumetric Ultrasound for Trauma Diagnosis

Computer-assisted 3D ultrasound methods for probe placement, kidney detection, and segmentation in abdominal trauma diagnosis, translated into patents and peer-reviewed research.

  • 3D ultrasound
  • Shape modeling
  • Segmentation

Experience

Research depth.
Production judgment.

A career connecting advanced perception research with dependable, deployable software.

July 6, 2026 — Present

Principal Research Developer, Computer Vision

Autodesk

Researching and developing applications of artificial intelligence, vision-language models, and computer vision.

2025 — Jul 2026

Lead AI Engineer, Perception & Sensor Fusion

Zadar Labs

Architected radar perception, camera-radar fusion, motion modeling, and real-time tracking systems for commercial sensing applications.

2020 — 2025

Co-Founder & Chief Technology Officer

EAIGLE

Built an AI company from research to commercial products; owned platform architecture and led a 15-person multidisciplinary team across enterprise deployments.

2018 — 2020

Senior Research Engineer, Computer Vision & ML

Huawei Technologies Canada

Developed deep-learning methods for object detection, tracking, and advanced imaging on mobile and embedded platforms.

2017 — 2018

Software Developer II

Seiko Epson Corporation

Developed robotic vision, object recognition, and gesture-tracking software for precision automation.

2016 — 2017

Research & Development Engineer

Magna International

Designed perception and sensor-fusion algorithms for automated parking and collision avoidance.

Technical practice

Full-stack AI expertise.

From mathematical modeling and training systems to optimized runtime and product integration.

A

AI & Machine Learning

  • Deep learning & model architecture
  • Supervised, semi-supervised & transfer learning
  • Transformers, VLMs, LLM applications & RAG
  • Training, optimization & evaluation pipelines
B

Computer Vision & Perception

  • Detection, segmentation & classification
  • Multi-object tracking & re-identification
  • 3D point clouds, radar & sensor fusion
  • OCR, keypoints, calibration & temporal modeling
C

Software & Systems

  • Python, modern C++, C#, JavaScript & SQL
  • PyTorch, TensorFlow, OpenCV & scikit-learn
  • ONNX, TensorRT & GPU inference
  • Modular architecture, CI/CD, edge & cloud

Research & invention

Ideas made measurable.

Ph.D. in Electrical & Computer Engineering from the University of Toronto, with peer-reviewed work in medical imaging, signal processing, and machine learning.

2023

FedCMR: A Library for Federated Continual Model Refinement

FLSys Workshop, MLSys

2018

Kidney Detection in 3-D Ultrasound Imagery via Shape-to-Volume Registration

IEEE Journal of Biomedical and Health Informatics

2016

An Automated Approach for Kidney Segmentation in 3D Ultrasound Images

IEEE Journal of Biomedical and Health Informatics

2016

Computer-Assisted 3D Ultrasound Probe Placement for Emergency Healthcare

IEEE Transactions on Industrial Informatics

View complete Google Scholar profile

Education

Ph.D. Electrical & Computer Engineering University of Toronto

M.A.Sc. Electrical & Computer Engineering Toronto Metropolitan University

M.Sc. Biomedical Engineering K. N. Toosi University of Technology

B.A.Sc. Electrical & Computer Engineering University of Tehran