Soroush Jaberi
CV (PDF)
Soroush Jaberi

Write

jaberi.soroush@gmail.com

Karaj, Iran · UTC+3:30

Elsewhere

About

I like results I can check, and I build things so that other people can check them too.

I am an AI researcher and engineer based in Karaj, Iran. I am completing an M.Sc. in Artificial Intelligence at Karaj Islamic Azad University with a GPA of 19.58 / 20 (4.0 / 4.0), where I also took my B.Sc. in Computer Engineering. My thesis is on classifying sentiment in Persian text by combining transformer models with lexicon-based methods. I expect to defend it in January 2027.

My recent work is about evidence: what a language model does when a retrieved document disagrees with what it knows, and what it costs to let the model reason about it. Before that I built document question answering with retrieval, sentiment classifiers for reviews, including Persian ones, and U-Net pipelines that segment tumors in CT scans. I am also writing a systematic mapping study of deep learning for biomedical image segmentation.

I have been a teaching assistant for three university courses, and I now teach in the open through The AI Path, a free week-by-week course that goes from Python to AI agents.

Download CV Get in touch GitHub

Soroush Jaberi, smiling, seated at a wooden table in a café
Soroush Jaberi سروش جابری

In frame

Soroush Jaberi standing outdoors on a sunny street, in a light sweater and tie
Soroush Jaberi seated at a table by a window, hands folded
Soroush Jaberi smiling, seated at a wooden table in a café
Soroush Jaberi in a cap and hoodie in front of bookshelves
Soroush Jaberi in sunglasses, lit by low sun in a café window

How I work

The six habits I practice, and teach every week in The AI Path.

  1. Understand the problem before you type.
  2. Take small steps and run often.
  3. Predict, then check.
  4. Treat errors as information.
  5. Verify, don’t trust.
  6. Leave it reproducible.

Teaching

  1. 2024 – 2025

    Teaching Assistant, Machine Learning (graduate level)

    Karaj Islamic Azad University

    • Assisted in designing and evaluating projects on supervised and unsupervised learning and model evaluation.
    • Supported graduate students in implementing machine learning workflows, and led Q&A sessions bridging theory and practice.

  2. 2023 – 2025

    Teaching Assistant, Data Mining

    Karaj Islamic Azad University

    • Worked with the course instructor to design and evaluate assignments on data preprocessing, feature selection, classification, clustering and pattern discovery.
    • Helped students bridge the gap between theoretical concepts and real-world applications in AI and analytics.

  3. 2022 – 2023

    Teaching Assistant, Algorithm Design

    Karaj Islamic Azad University

    • Assisted in developing and assessing problem sets requiring efficient algorithm design and rigorous complexity analysis.

Education

  1. 2023 – present

    M.Sc. in Artificial Intelligence

    Karaj Islamic Azad University

    • GPA 19.58 / 20 (4.0 / 4.0 by course-by-course conversion)
    • Thesis defense expected January 2027
    • Supervisor: Dr. Amineh Amini. Co-advisor: Dr. Hadi Saboohi
    • Coursework: Machine Learning (20 / 20), Neural Networks & Deep Learning (20 / 20), Image Processing (19.5 / 20)

  2. 2019 – 2023

    B.Sc. in Computer Engineering

    Karaj Islamic Azad University

    • GPA 17.97 / 20 (3.74 / 4.0 by course-by-course conversion)
    • 3.81 / 4.0 over the last two years
    • Coursework: Artificial Intelligence (20 / 20), Data Mining (20 / 20), Algorithm Design (20 / 20), Computer Graphics (19.5 / 20)

Research and publications

Tools

What I reach for. The work pages show them in use.

Programming
Python, SQL, Java, JavaScript, HTML/CSS
Machine learning & deep learning
PyTorch, TensorFlow, scikit-learn, NumPy, pandas, OpenCV, NLTK, XGBoost, LightGBM
NLP, LLMs & retrieval
Transformers (ParsBERT, BERT), Retrieval-augmented generation, LangChain, ChromaDB, Embedding-based retrieval, vLLM
Tools
Git/GitHub, Linux (Ubuntu/Debian), Anaconda, Jupyter, Google Colab, PyCharm, VirtualBox/VMware, Minitab
Spoken
Persian — native; English — advanced (academic reading and technical writing)