cv
Basics
Name | Alireza Salehy |
Label | Master student |
alireza99salehy@gmail.com | |
Url | https://alirezasalehy.github.io/ |
Summary | Master's student in Artificial Intelligence at the University of Tehran, focusing on vision-language models, few-shot anomaly detection, and robust model adaptation. |
Work
-
2021.01 - 2023.09 Firmware Engineer
NOTRIX (Remote)
Developed firmware for smart home devices and their drivers for main controller integration.
- IoT Firmware Development
- Device Driver Integration
Education
-
2022.09 - 2025.02 Tehran, Iran
MSc
University of Tehran
Artificial Intelligence
- Interpretable and Trustworthy AI (20/20)
- Deep Generative Models (19.8/20)
- Design and Analysis of DNNs (19.85/20)
- Deep Learning with Vision and Language Applications (19.83/20)
- Statistical Inference (16.3/20)
- Data Analytics and Visualization (17.93/20)
-
2017.09 - 2022.06 Tehran, Iran
BSc
Amirkabir University of Technology
Computer Engineering
- Probability and Statistics (20/20)
- Applied Linear Algebra (17.2/20)
- Data Structure and Algorithms (17.25/20)
- Assembly and Microprocessors (20/20)
- Operating Systems (17.2/20)
- Computer Networks (18/20)
- Internet of Things (18.25/20)
Awards
- 2022.07.01
National Master Entrance Exam – AI Branch
Ministry of Science, Iran
Ranked 23rd out of 15,000 (top 0.2%) nationwide.
- 2017.07.01
National Bachelor Entrance Exam – Math & Physics Branch
Ministry of Science, Iran
Ranked 263rd out of 44,000 (top 0.6%) nationwide.
Publications
-
2025.03.01 Crane: Context-Guided Prompt Learning and Attention Refinement for Zero-Shot Anomaly Detection
Submitted to ICCV 2025
Proposes Crane, a CLIP-based architecture enhancing zero-shot anomaly detection through context-aware prompting and attention refinement.
Skills
Machine Learning | |
Vision-Language Models | |
Contrastive Learning | |
Zero-shot Learning | |
Few-shot Learning | |
Anomaly Detection |
Embedded Systems | |
STM32 | |
ESP32 | |
FreeRTOS | |
Embedded C/C++ | |
IoT |
Languages
English | |
Fluent |
Persian | |
Native |
Interests
AI Research | |
Vision-Language Models | |
Robustness | |
Fairness in AI |
Embedded Systems | |
IoT | |
Low-power Design | |
Firmware Development |
References
Dr. Mohammad Sabokrou | |
Associate Professor, Machine Learning Lab, Okinawa Institute of Science and Technology – mohammad.sabokrou@oist.jp |
Dr. Reshad Hosseini | |
Associate Professor, University of Tehran – reshad.hosseini@ut.ac.ir |
Projects
- 2024.08 - 2024.10
Crane – Zero-Shot Anomaly Detection with CLIP
Developed a state-of-the-art anomaly detection method using prompt learning and spatial attention refinement on CLIP.
- Prompt Learning
- Attention Weighting
- Local-Global Embedding Fusion
- 2021.01 - 2021.06
WAVRecorder C++ Library
Developed a C++ library for efficient voice recording using embedded ADC and SD card.
- Low-Level Programming
- Embedded Audio Processing