cv

Basics

Name Alireza Heidari
Label M.Sc. Student in Computing Science
Email alirezaheidari.cs@gmail.com
Url https://deepmancer.github.io/
Summary Thesis-based M.Sc. student at Simon Fraser University working on computer vision, computer graphics, generative modeling, self-supervised learning, machine learning theory, and representation learning.

Work

  • 2026.02 - Present

    Vancouver, Canada

    Machine Learning Engineer, VanityAI
    MARZ - Monsters Aliens Robots Zombies
    Part-time research and engineering role on machine-learning components for VanityAI.
    • Supported the revival of VanityAI by contributing to core machine-learning research and engineering work.
  • 2022.01 - 2023.03

    Tehran, Iran

    Mid-Level Data Scientist, Artificial Intelligence - Maps
    Tapsi
    Worked on ETA prediction, map matching, routing, and data science infrastructure for a large ride-hailing platform.
    • Centralized ETA modules across backend and data science, aligning simulation and production metrics and contributing to a 1% revenue increase.
    • Prototyped a graph-based ETA prediction system using road-segment graphs and Graph Neural Networks.
    • Improved OSRM-based preprocessing for Hidden Markov Model map matching, reducing matching errors by nearly 80%.
    • Built an automated model retraining and deployment pipeline for the ETA framework.
  • 2021.02 - 2022.01

    Tehran, Iran

    Mid-Level Software Engineer, Backend - Rides
    Tapsi
    Worked on ride matching, dispatch, microservices, and observability for the core rides platform.
    • Collaborated on a flow-optimized driver-passenger matching algorithm that increased revenue by nearly 2% and reduced average wait time by 8%.
    • Refactored three microservices with query optimization, indexing, caching, and gRPC communication, reducing average response times by about 15%.
    • Developed Grafana and Metabase monitoring dashboards for production microservices.

Research

  • 2025.01 - Present

    Burnaby, Canada

    Research Assistant in 3D Computer Vision
    GrUVi Lab, Simon Fraser University
    Supervised by Dr. Ali Mahdavi-Amiri.
    • Developing a 3D hair modeling framework that generates realistic hair from a single image.
    • Training on diverse synthetic hairstyles while prioritizing generalization to real portrait domains.
  • 2023.07 - 2024.08

    Tehran, Iran

    Research Assistant in Multimodal Learning
    Sharif University of Technology
    Supervised by Dr. Mahdieh Soleymani Baghshah.
    • Studied inconsistencies in vision-language models across different levels of label abstraction.
    • Integrated semantic hierarchical label knowledge during fine-tuning to improve class separation across granularities.
  • 2023.01 - 2023.09

    Tehran, Iran

    Research Assistant in Machine Learning Theory
    Sharif University of Technology
    Supervised by Dr. Abolfazl Motahari and Dr. Amir Najafi.
    • Combined self-supervised learning and distributionally robust learning to use out-of-domain unlabeled samples.
    • Helped develop polynomial-time algorithms with improved theoretical generalization bounds over ERM baselines.

Education

  • 2025.01 - Present

    Burnaby, Canada

    M.Sc. (Thesis-Based)
    Simon Fraser University
    Computing Science
    • GPA: 4.08/4.00
    • Supervisor: Dr. Ali Mahdavi-Amiri
  • 2019.09 - 2024.11

    Tehran, Iran

    B.Sc.
    Sharif University of Technology
    Computer Engineering
    • GPA of units passed: 17.93/20
    • Thesis: Enhancing Vision-Language Models' Classification Performance with Hierarchical Semantic Labels
    • Supervisor: Dr. Mahdieh Soleymani Baghshah
    • Graduate-level courses: Deep Learning (20/20), Machine Learning (20/20)

Publications

  • 2026
    HairPort: In-context 3D-Aware Hair Import and Transfer for Images
    ACM SIGGRAPH, 2026
    A 3D-aware in-context hair import and transfer method for images.
  • 2024
    Unlabeled Out-of-Domain Data Improves Generalization
    International Conference on Learning Representations (ICLR), 2024
    Spotlight presentation.

Projects

  • - Present
    Vision-Language Models Toolbox
    A PyTorch library for multimodal research with vision-language models such as CLIP and DINO-V2, supporting fine-tuning strategies including contrastive and soft-prompt tuning.
  • - Present
    MedSegDiff
    PyTorch reimplementation of MedSegDiff for tumor detection.
  • - Present
    Zero-Shot Object Detection
    Zero-shot object detection pipeline combining CLIP with Faster R-CNN region proposals.
  • - Present
    Bootstrap Your Own Latent (BYOL)
    Fine-tuning BYOL models on STL10 with a pretrained ResNet backbone.
  • - Present
    Generative Models
    Comparative study of VAEs, GANs, and DDPMs on Fashion MNIST.
  • - Present
    Full-Stack Food Delivery Platform
    A FastAPI and Vue.js food delivery application built on an event-driven distributed microservices architecture.
  • - Present
    FastAPI JWT Authentication
    Python authentication middleware for FastAPI with Redis integration.

Teaching

  • 2025
    Special Topics in Artificial Intelligence
    Simon Fraser University
    Lectured by Dr. Nick Vincent
  • 2024 and 2023
    Deep Learning (Graduate Course, x2)
    Sharif University of Technology
    Lectured by Dr. Beigy and Dr. Soleymani Baghshah
  • 2024
    Deep Generative Models (Graduate Course)
    Sharif University of Technology
    Lectured by Dr. Beigy
  • 2023 and 2022
    Machine Learning (Graduate Course, x2)
    Sharif University of Technology
    Lectured by Dr. Sharifi Zarchi and Dr. Motahari
  • 2023
    Advanced Information Retrieval
    Sharif University of Technology
    Lectured by Dr. Beigy
  • 2023
    Artificial Intelligence
    Sharif University of Technology
    Lectured by Dr. Soleymani Baghshah and Dr. Rohban
  • 2023
    Intelligent Analysis of Biomedical Images (Graduate Course)
    Sharif University of Technology
    Lectured by Dr. Rohban

Awards

  • 2019
    Silver Medal
    International Olympiad on Astronomy and Astrophysics
    13th IOAA, Hungary.
  • 2018
    Gold Medal (1st Rank)
    Iran's National Olympiad on Astronomy and Astrophysics
    Ranked first nationally.
  • 2018
    Elite Recognition
    Iran's National Elites Foundation
    Recognized by Iran's National Elites Foundation since 2018.

Skills

Deep Learning
PyTorch
Hugging Face Transformers
RAPIDS
TensorFlow
Keras
PyTorch Geometric
Machine Learning
Spark MLlib
Scikit-Learn
CatBoost
XGBoost
PySpark
Pandas
NumPy
SciPy
Dashboarding and Visualization
Metabase
Grafana
Tableau
Power BI
Dash
Plotly
Matplotlib
Seaborn
End-to-End Machine Learning Workflow
MLflow
Metaflow
Prefect
Weights & Biases
Luigi
DBMS
pgAdmin
RedisInsight
MongoDB Compass
PostgreSQL
MongoDB
Redis
Hazelcast
MySQL
Programming Languages
Python
Django
Node.js
Java
Go
C++
C
R
LaTeX

Languages

Persian
Native speaker
English
IELTS Academic 7.0/9.0, DET 140/160

Interests

Research Interests
Computer Vision
Computer Graphics
Generative Modeling
Self-Supervised Learning
Machine Learning Theory
Representation Learning
Activities
Piano
Observational Astronomy
Tennis
Swimming
Table Tennis