News

  • SEGA accepted at NeurIPS 2026 🎉

  • Awarded the Graduate Scholarship in Science and Technology ($15,000 CAD), Mount Sinai Hospital.

  • PhotoQuilt preprint released on arXiv.

  • SEGA preprint released on arXiv.

  • Paper on nursing-activity recognition published in Machine Learning: Health.

  • Oral presentation of our mmWave radar work at IEEE EMBC 2025.

  • Survey on indoor localization published in Computer Networks.

  • FUVT published in Signal, Image and Video Processing.

  • Awarded the EPIC Doctoral Award ($10,000 CAD), Emerging and Pandemic Infections Consortium.

Selected publications

Google Scholar
  1. NeurIPS 2026
    SEGA: Spectral-Energy Guided Attention for Resolution Extrapolation in Diffusion Transformers

    Javad Rajabi, Kimia Shaban, Koorosh Roohi, David B. Lindell, and Babak Taati

    Advances in Neural Information Processing Systems (NeurIPS), 2026

    arXiv project code

  2. Preprint 2026
    PhotoQuilt: Training-Free Arbitrary-Resolution Photomosaics via Bootstrapped Tiled Denoising

    Koorosh Roohi*, Javad Rajabi*, Andrew Fleet, and Babak Taati

    arXiv preprint arXiv:2606.30968, 2026

    arXiv PDF (HQ) project code

  3. MLHealth 2026
    Deep learning-based recognition of simulated nursing activities for infection risk assessment

    Koorosh Roohi and Atena Roshan Fekr

    Machine Learning: Health, 2026

    paper

  4. EMBC 2025
    Using mmWave radar and deep learning to classify caregiver activities for infection prevention

    Koorosh Roohi and Atena Roshan Fekr

    IEEE Engineering in Medicine and Biology Conference (EMBC), 2025 Oral

    paper

  5. ComNet 2025
    A comparative analysis of indoor localization technologies: a review of the literature

    Koorosh Roohi and Atena Roshan Fekr

    Computer Networks, 2025 Survey

    paper

  6. SIVP 2025
    FUVT: a deep few-shot unsupervised learning-based video-to-video translation scheme using Kalman filtering and relativistic GAN

    Koorosh Roohi, Alireza Esmaeilzehi, and M. Omair Ahmad

    Signal, Image and Video Processing, 2025

    paper

  7. Under review
    Real-time caregiver activity recognition with potential for infection prevention in hospitals

    Koorosh Roohi and Atena Roshan Fekr

    Under review at Smart Health

* Equal contribution.

Experience

  • 2023 — 2026

    AI Manager · Masterly Inc., Toronto

    AI products for organizations; cloud deployment on AWS.

    • Led end-to-end development of Masterly Forms, a collaborative platform with AI-assisted form design and completion.
    • Built Masterly Earendel, a vision–language-model-based video analysis pipeline for event detection and alerting in an agentic architecture.
    • Led development of Masterly FinSuite, an LLM-driven financial portal for reporting automation and decision support.
    • Built production infrastructure on AWS (SageMaker, Bedrock, Lambda, EC2); supervised a team of five developers and shipped modules in PyTorch and .NET.
  • 2022

    Computer Vision Engineer · Crouse Company, R&D, Tehran

    Quality inspection for automotive displays.

    OpenCV-based defect-detection pipeline; Isolation Forest and Tesseract OCR for pixel-level anomaly screening.

  • 2019 — 2022

    Client Developer · Quiz of Kings Studio, Tehran

    Unity client for a large-scale mobile title.

    Client-side feature work and migration of legacy code from Corona to Unity.

Background — briefly

Education

  • Ph.D. candidate, Biomedical Engineering, University of Toronto 2023 onward · expected 2028 Dissertation learning and multi-modal sensing in clinical settings, including 3D data, sequential modeling, and system integration.
  • B.Sc., Computer Engineering, Amirkabir University of Technology Tehran Polytechnic 2023

Teaching

University of Toronto · Department of Computer Science (tutorial and lab TA)

  • CSC 209 — System Programming
  • CSC 385 — Microprocessor Systems
  • CSC 207 — Software Design