Research Interests

  • NLP
  • Deep Learning
  • Machine Learning
  • Trustworthy AI

Education

Publications

  1. Fine-Tuning OPT, Mistral, and LLaMA for Paraphrase-Robust Clinical Semantic Textual Similarity using Deep Metric Learning. • First Author Ongoing research (expected January 2026)
  2. Trends and Shifts in Groundwater Levels Across Metropolitan France. • First Author Under review at Scientific Reports Access upon request
  3. FaBERT: Pre-training BERT on Persian Blogs. • Second Author Proceedings of the NAACL Workshop on Noisy User-generated Text (W-NUT), 2025 ACL
  4. hmBlogs: A Comprehensive Corpus and Benchmarking Study for Persian Language Embeddings. • Fourth Author Under review at SN Computer Science Access upon request
  5. Intra Class Feature Learning and Supervised Triplet Sampling for Deep Metric Learning     • Third Author Under review at IKT 2026 Access upon request
  6. Evaluating LLMs in Persian News Summarization. • Fourth Author 2024 15th International Conference on Information and Knowledge Technology (IKT) IEEE
  7. FarsEval-PKBETS: A New Diverse Benchmark for Evaluating Persian Large Language Models. • Fourteenth Author arXiv:2504.14690 arXiv
  8. Cluster Sampling: A Cluster-Driven Sampling Strategy for Deep Metric Learning. • Third Author 2024 14th International Conference on Computer and Knowledge Engineering (ICCKE) IEEE

Research Experience

France Underground Water Level Assessment — BRGM

Conducted statistical analysis of underground water levels from piezometer networks across Metropolitan France in collaboration with BRGM (French Geological Survey) to support sustainable water resource management. Designed and implemented data pipelines for long-term hydrological records, developed visualizations, and interpreted basin-level trends and anomalies to inform decision-making.

Access upon request

Supervisor: Dr. Ahmad Ali Abin

FarsEval: Benchmarking Large Language Models for Persian NLP

–

Developing datasets and evaluation metrics for assessing LLM performance in summarization, analogy testing, formality style transfer, and other Persian NLP tasks.

arXiv

Supervisor: Dr. Mehrnoush Shamsfard

Graduate Research — NLP Lab, Shahid Beheshti University

–

Improved clinical Semantic Textual Similarity (STS) using metric learning and MeSH ontology. Developed paraphrase-robust language model training strategies to enhance model robustness while managing the trade-off with generalization.

Proposal (EN) Thesis (FA)

Supervisor: Dr. Mehrnoush Shamsfard

Research Assistant — Iranian Research Organization for Science and Technology (IROST)

–

Researched different rPPG (Remote Photoplethysmography) signal extraction methods using videos from the skin at different light wavelengths to map blood flow below the skin, in collaboration with The Color and Visual Computing Laboratory (Colourlab) at the Norwegian University of Science and Technology (NTNU).

Supervisor: Dr. Vahidreza Nafisi

Undergraduate Research — Amirkabir University of Technology

–

Estimated the heart rate and mapped blood flow below the skin using videos from facial regions with a non-invasive rPPG method. Implemented Region of Interest (ROI) detection (using neural networks, thresholding, and manual selection), extracted signals via average pixel intensity and Independent Component Analysis (ICA), and applied the Fourier Transform to identify the dominant frequency corresponding to the heart rate.

Thesis (FA)

Supervisor: Dr. Vahidreza Nafisi

Teaching

Teaching Assistant — Department of Computer Science & Engineering, Shahid Beheshti University

–
Course: Introduction to Machine Learning
Instructor: Prof. Hamed Malek · Semester: Spring 2024

Teaching Assistant — Department of Computer Engineering, Amirkabir University of Technology

–
Course: Computer Architecture
Instructor: Prof. Hamed Farbeh · Semester: Spring 2022

Teaching Assistant — Department of Computer Engineering, Amirkabir University of Technology

–
Course: Microprocessors and Assembly Language
Instructor: Prof. Hamed Farbeh · Semester: Fall 2019

Selected Course Projects

  • Robust FAQ Question Answering using Metric Learning (Angular Loss) and SPARQL-Based Ontology Querying. Ontology & Knowledge Engineering GitHub 📄Report
  • Evaluating Fairness, Backdoor Attack, and Out-Of-Distribution Detection in ResNet18 Image Classification. Interpretable AI GitHub 📄Report
  • Model Interpretation using SHAP for Regression (MLP) and LIME for Image Classification (MobileNet v2). Interpretable AI GitHub 📄Report
  • Enhancing Robustness and Generalization in ResNet18 Classification with Angular Loss on Limited Training Data (20% of CIFAR-10 dataset) Interpretable AI GitHub 📄Report
  • Question Answering Using Cosine Similarity Between Input Questions and Constructed Sentences over RDF Ontologies (S-P-O). Ontology & Knowledge Engineering GitHub 📄Report
  • Comparing Performance of Contextual and Static Embeddings in Analogical Reasoning Tasks. Natural Language Processing GitHub
  • Image Colorization (CNN). Deep Learning GitHub
  • Tata Steel Stock Price Prediction (HMM). Machine Learning GitHub
  • Text Classification on WELFake — BERT & RoBERTa. Deep Learning GitHub
  • Sign Language Image Classification — MLP. Deep Learning GitHub
  • Human Activity Recognition — CNN-LSTM. Neural Networks GitHub
  • Google Stock Price Prediction — GRU/LSTM. Deep Learning GitHub
  • Earthquake Prediction in Iran. Machine Learning GitHub
  • Baseline Wander Removal in ECG — FIR, IIR, EMD, Moving Avg, Wavelet. Digital Signal Processing
  • Object Recognition: Humans vs CNNs (Kay et al., 2008). Neuromatch Academy

Selected Coursework

  • Interpretable AI (19.4/20) — Visiting, University of Tehran — 📘ABET
  • Natural Language Processing (17/20)
  • Deep Learning (17.3/20)
  • Machine Learning (16.8/20)
  • Knowledge & Ontology Engineering (19.25/20)
  • Principles of Algorithms (20/20) — Visiting, K. N. Toosi University of Technology
  • Data Structures and Algorithms (17/20)
  • Digital Signal Processing (19/20)
  • Pattern Recognition (16.75/20)

Professional Development

Honors & Awards

  • Top 1% in Iran’s Nationwide University Entrance Exam (Konkour) for Engineering — Rank 963 (Region 1), 2016.
  • Member, National Organization for Development of Exceptional Talents (NODET), Tehran (2012–2016).

References