Academic & research experience
Curriculum vitae
Research experience, publications, and technical expertise.
Research focus
Machine learning, vision & health.
I develop methods that help AI models adapt to new settings with stronger performance and lower computational costs. My work spans parameter-efficient fine-tuning, physics-informed modeling, and model evaluation, including settings with limited labeled data and noisy signals. I apply these approaches to digital health, computer vision, acoustics, and sensing.
Research & contributions →Education
Academic training
- PhD in Computer Engineering
Texas A&M University
September 2022 – December 2026 - MS in Electrical Engineering
Sharif University of Technology
September 2018 – February 2021 - BS in Electrical Engineering
University of Tabriz
September 2014 – September 2018
Research experience
Graduate Research Assistant · Texas A&M University
September 2022 – December 2026
- Developed parameter-efficient fine-tuning (PEFT) methods for Transformer foundation models, adapting pretrained representations to new classification tasks through distribution-aware modules.
- Built a few-shot histopathology classification pipeline combining vision-language models, attention-based multiple instance learning, and visual and textual adaptation.
- Improved mean balanced accuracy from 43.5% to 78.1% in a four-shot whole-slide classification study using subject-disjoint splits and repeated evaluation.
- Contributed to a modular evaluation pipeline for image classification, semantic segmentation, and object detection, comparing pretrained computer vision models and adaptation strategies.
- Developed physics-informed neural networks for physiological time-series prediction; the collaborative blood-pressure study reduced ground-truth training requirements by an average factor of 15.
- Benchmarked cross-domain transfer learning, finding a 3-percentage-point classification accuracy improvement from image-pretrained models over audio-pretrained models.
Graduate Student Researcher · Sharif University of Technology
July 2019 – February 2021
- Designed an integrated ECG/EDA wearable sensor, from circuitry and Bluetooth Low Energy electronics to machine-learning algorithms.
- Achieved 94% stress-detection accuracy in an 18-participant study and threefold battery life relative to the evaluated alternatives.
Publications
Paper details, publication status, and available resources are listed in the HTML publication record.
Read publications →Teaching & service
Mentoring, guest lectures, peer reviewing, session chairing, and conference travel awards.
Background & service →Updated October 24, 2026. For an alternative format, contact me by email.