Selected Work
Research & Projects
A selection of research and hands-on projects spanning EEG, multimodal NLP, and Persian speech processing. Click any project to read more.
Multimodal Sentiment Analysis with EEG & Language
- Investigating cross-cultural neural correlates of language and sentiment using EEG from Persian and English speakers.
- Developing multimodal sentiment models that integrate EEG signals with textual representations.
- Building EEG feature-extraction pipelines with MNE-Python for multimodal machine learning.
- Implementing cross-lingual representation learning across Persian and English EEG–text data.
- Contributed to PARSA-Bench, benchmarking large audio-language models for Persian.
Selected Projects
ASR-Based Gender & Accent Classification
End-to-end Persian speech pipeline and Wav2Vec2 fine-tuning for ASR and speaker-attribute analysis.
Neural Decoding & Psychophysics Analyses
Psychometrics, spike sorting from scratch, and EEG decoding (MVPA, time-generalization, RSA vs. CORnet-S).
Transfer-Learning-Based Food Recognition
Fine-tuned convolutional neural networks for food-image recognition — B.Sc. thesis (ranked 1st of 96).
Toolkit
Python
PyTorch
TensorFlow
scikit-learn
Hugging Face
LangChain
LangGraph
FAISS
librosa
fairseq
ESPnet
SpeechBrain
MNE-Python
EEGLAB
Tensorpac
NumPy
pandas
Git
LaTeX