Curriculum Vitae

York University – M.Sc in Computer Science

• Supervised by supervised by Dr. Marcus A. Brubaker
• Thesis: Scale Ambiguity in Generative Novel View Synthesis
• GPA: A+

Sharif University of Technology – B.Sc. Computer Engineering

• Supervised by Dr. Mohammad Hossein Rohban.
• Thesis: Improve cell segmentation using self-supervised frameworks (SimCLR, MoCo, SimSiam)
• GPA: 18.53 (20.00 scale)

Machine Learning Intern Vector Institute

• Advised by Dr. Arash Afkanpour
• Working on Generative Self-Supervised Learning.
• Adding semantically consistent generative image augmentations with Instance-Conditioned GAN and Stable Diffusion UnClip to various self-supervised methods.
• Enhanced the generalization and robustness of self-supervised representations in all cases.

Graduate Research Assistant at Computational Vision and Imaging Lab

• Advised by Dr. Marcus A. Brubaker
• Working on Scale Ambiguity in Generative Novel View Synthesis (GNVS).
• Introduced Scale Ambiguity as a challenge in single view GNVS with various experiments.
• Proposed to optimize scales per training scene in a Multi-View Diffusion Model used for Novel View Synthesis.
• Introduced a new metric to based on Optical Flow to measure the effectiveness of optimizing scale with denoising loss.

Summer Intern at VITA Lab, EPFL

• Supervised by Dr. Alexandre Alahi
• Conducted a literature review on density estimation techniques and their applications on human trajectory data.
• Used masked autoregressive flow to find natural adversarial examples to test the reliability of human trajectory predictors.
• Adversarially trained LSTM based predictors and reduced the collision rate up to 35%in the case of an adversarial attack on test data.
• Project page: [Translated page]

Research Assistant at Medical Image Analysis Group, Department of Computer Engineering, Sharif University of Technology

• Advised by Dr. Mohammad Hossein Rohban
• Used unsupervised learning frameworks (simCLR,MoCo,SimSiam) to train U-net encoder with unannotated cell images.
• Improved mean average precision (mAP) after fine-tuning with annotated ones up to 8%.

Machine Learning Intern at Sinaweb Company

• Extracting lexical, structural, and syntax features.
• Proposed a regression model to fuse features and predict writing style.
• Implemented an outlier detection model to find possible plagiarised segments.

• Flight PS752 Commemorative Scholarship, Global Affairs Canada (29,000 CAD) (2024)
• Vector Scholarship in AI (17,500 CAD) (2022)
• VISTA Program Master's Scholarship (20,000 CAD) (2022)
• York Graduate Scholarship (6000 CAD) (2022)

• Ranked 125^th among more than 150,000 participants in Nationwide Universities Entrance Exam (B.Sc.). (2017)
• National Elite Foundation Fellowship.

• Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization(link)
• Neural Networks and Deep Learning(link)

• Deep Unsupervised Learning (CS294), UC Berkeley. (link) CNNs for Visual Recognition (CS231), Stanford University. (link)

• Python, Java, C/C++.

• Pytorch, Pytorch lightning, Keras, NumPy, scikit-learn, SciPy, Pandas.

• Multi-GPU Training, Distributed Training, SLURM, Conda, Github’s CI/CD, LaTeX.

• SQL, HTML/CSS.