AI Safety & Reliability Research

Hamidreza
Yaghoubi

Ph.D. Student · UMD GAMMA Lab

I'm a Ph.D. student in Computer Engineering at University of Maryland, working in GAMMA Lab under Prof. Ming Lin. My research focuses on robust and reliable machine learning, AI safety, and spurious correlations in vision systems.

Hamidreza Yaghoubi
01

About

I'm a Ph.D. student in Computer Engineering at the University of Maryland, College Park, working in GAMMA Lab under the supervision of Professor Ming Lin.

Previously, I completed my B.Sc. in Computer Engineering from Sharif University of Technology, where I worked with Professor Mahdieh Soleymani and Professor Mohammad Hossein Rohban.

My research interests include robust and reliable machine learning, interpretability and faithfulness of vision–language models, spurious correlations and shortcut learning, and AI safety. A recurring theme in my work is understanding when models appear accurate on average but behave unexpectedly in important or edge cases.

02

Publications & Preprints

01
Recti-Q: Feature-Space Rectification for Out-of-Distribution-Robust Quantized Perception in Edge Robotics
Hamidreza Yaghoubi*, Parastoo Pilevar*, and Ming C. Lin
02
PolySona: Parameter-Efficient Driving Style Modeling for Trajectory Prediction
Laura Zheng, Hamidreza Yaghoubi, Tony Wu, Tianyi Zhou, and Ming C. Lin
03
Laura Zheng*, Hamidreza Yaghoubi*, Tony Wu, Sandeep Thalapanane, Tianyi Zhou, and Ming C. Lin
04
Fahimeh Hosseini Noohdani, Parsa Hosseini*, Aryan Yazdan Parast*, Hamidreza Yaghoubi, Mahdieh Soleymani Baghshah
05
Mahdi Ghaznavi, Hesam Asadollahzadeh, Hamidreza Yaghoubi, Fahimeh Hosseini Noohdani, Mohammad Hossein Rohban, Mahdieh Soleymani Baghshah

* Equal contribution

03

Work Experience

Data Scientist
TAPSI — Iranian Version of Uber
Feb 2021 – Feb 2023

TAPSI is an Iranian ride-hailing company similar to Uber, and stands as one of Iran's largest and most technologically advanced companies. Notable projects I contributed to include:

Estimated Time of Arrival (ETA): Investigated and developed traffic prediction models using PyTorch, PySpark, and Luigi. Added lightweight post-processing that reduced absolute error by 3%, increased completed rides by 1.5%, and eliminated a data-prep bottleneck enabling 60× faster inference.
Location Search Engine: Processed raw data and built an in-house search engine with Elasticsearch, including geodata ingestion and cleaning pipelines. Defined offline and online quality metrics aligned with marketplace KPIs.
GPS Denoising: Applied noise reduction techniques to enhance GPS data quality and reliability for location-based services.
Destination Suggestion: Built a Gaussian Mixture Model-based destination recommender using user ride history and behavioral patterns.
04

Skills

Programming
Python Java C C++ SQL Bash
Software
PyTorch TensorFlow Keras PySpark Pandas NumPy Matplotlib Seaborn Luigi
Languages
English Persian
05

Honors & Awards

2025
Fellowship — UMD ECE Summer Research Fellowship
University of Maryland, College Park · Maryland, USA
2019
Bronze Medal — 13th International Olympiad on Astronomy and Astrophysics
Keszthely, Hungary
2018
Gold Medal — 14th National Olympiad on Astronomy and Astrophysics
Tehran, Iran
2015
Winner — 1st Tehran and Alborz Province Inventions Festival
Tehran, Iran
06

Teaching Experience

Teaching Assistant
University of Maryland, College Park
Fall 2024 – Present
Computer Organization
Teaching Assistant
Sharif University of Technology
2020 – 2024
Graduate: Deep Learning, Machine Learning, Intelligent Analysis of Biomedical Images
Undergraduate: Artificial Intelligence, Advanced Programming
Teacher
Top High Schools in Iran
2018 – 2022
Instructing and mentoring students for National and International Olympiads in Astronomy and Astrophysics.
07

Academic Service

Reviewer Invited
NeurIPS 2026
Reviewer Invited
ECCV 2026
Reviewer Invited
CVPR 2026
Reviewer
ICRA 2025
Branding Team Member, Webelopers Web Programming Competition
2021
Practical Round Committee Member, 15th Asian-Pacific Astronomy Olympiad
2019