Kian Khooban · Workbench
25°CSET 350°C

unrolling the cutting mat...

simple version

Kian Khooban

Software engineer: full-stack apps, AI agents, database operations, AI hardware and ML research. Third-year CS at York University.

Hardware Research

Simulating analog AI accelerator chips at York.

ML Research

Medical image segmentation in PyTorch.

  • End-to-end PyTorch pipeline for medical image segmentation: YOLO trained on Google Cloud, human-in-the-loop labelling with CVAT and SAM2.
  • Caught a train/test leak by hashing 5,400 frames (68.9% of the test set duplicated training data) and rebuilt evaluation on a subject-level split.

Experience

Software engineering and database operations.

  • IT Assistant, Database Services, York University (Sep 2026 to now): patching production SQL Server clusters through zero-downtime rolling failovers, automating the cycle in PowerShell.
  • Software Engineer, Hoshmand Afarin (Apr 2025 to Apr 2026): React investor dashboard for 20 clients, MySQL schema turning live market feeds into per-investor returns.
  • Campus Partner, Perplexity (Sep 2025 to Jan 2026): live workshops for about 60 students, 30+ sign-ups.

About

Third-year CS at York University, Lassonde.

    Contact

    Get in touch.

    Resume

    The spec sheet.

    Projects

    Things that shipped.

    Colophon

    An interactive 3D workbench built with Next.js + React Three Fiber.

    Mat, layout, and everything else: design and code by Kian Khooban.