Ph.D. Candidate · University of Victoria

Nafiseh Izadyar

Working in Computer Graphics and Machine Learning, with a focus on CAD systems, boundary representation (B-Rep), and geometric modeling, and their role in representing and analyzing complex geometric structures.

What I Work On

LLM-Guided 3D Inference

Using Large Language Models to guide material inference and reasoning over 3D point clouds and CAD models, bridging language understanding with geometric data.

LLMs 3D Point Clouds Multimodal

Boundary Representation

Developing better formats and datasets for B-Rep modeling, including topology-aware meshing and half-edge data structures for CAD applications.

B-Rep STEP Files CAD

Geometric Representation Learning

Learning robust representations of 3D geometric data through multimodal models, combining visual features with structural and material properties.

Deep Learning Representation Learning 3D Data

Selected Works

Academic Background

Ph.D. in Computer Science
University of Victoria, Canada
Sept 2021 – Feb 2027 (expected)
Thesis: Advancing Geometric Representation and Material Inference for 3D CAD Models. Supervised by Dr. Teseo Schneider. Focus on LLMs, multimodal learning, and B-Rep modeling.
M.Sc. in Artificial Intelligence
Amirkabir University of Technology, Tehran
Sept 2016 – Feb 2019
Thesis: Visual Question Answering using Deep Learning, Computer Vision, NLP, and Reinforcement Learning. Advised by Dr. Ahmad Nickabadi.
B.Sc. in Computer Software Engineering
Shariaty Technical and Vocational University, Tehran
Sept 2011 – Sept 2015
Thesis: Fingerprint Recognition. Awarded Best Project of the Year.

Teaching and Professional Experience

Sessional Instructor — CSC 115

University of Victoria · May – August 2026

Instructor for CSC 115: Fundamentals of Programming II.

Teaching Assistant

University of Victoria · Sept 2021 – Present

Led tutorials and labs for undergraduate and graduate CS courses. Mentored students on ML, algorithms, and software engineering.

Research Assistant

Department of Computer Science, UVic · Sept 2021 – Present

Designed ML pipelines for large-scale CAD datasets, implemented evaluation metrics for CAD model validity, and worked with multimodal models combining geometric data with learned representations.

Innovation & Systems Co-op

Legislative Assembly of British Columbia · May – Aug 2023

Applied AI governance and deployment in a production government environment. Integrated AI tools with attention to reliability, performance, and stakeholder needs.

Tools & Expertise

ML & LLMs

PyTorch TensorFlow LLMs Multimodal Models Reinforcement Learning Representation Learning

CAD & Geometry

B-Rep STEP Files OpenCascade Geometric Modeling Half-Edge Structures

Systems & Languages

Python C++ CUDA SQL Linux Git Bash

Graphics & Vision

3D Data Processing OpenCV Point Clouds OpenAI Gym

Let's Connect

I'm always open to discussing research collaborations, speaking opportunities, or questions about my work in ML and computer graphics.