S. F. Mona Ebrahimi

| AI | Machine Learning | Data Science | NLP | Doctoral Researcher | Tampere University / Sharif University of Technology
πŸš€ Currently seeking a postdoctoral or research position. I'm especially drawn to real-world health problems β€” research that leaves the lab and actually helps people β€” and I'm open to a range of directions in machine learning, data science, and NLP. If you're hiring or know of a good fit, let's talk.

Oh, hello there! Welcome to my little corner of the world-wide web! Make yourself at home. I'd offer you a real cup of tea and a cookie, but unfortunately, my website only supports the digital kind. β˜•πŸͺ

  • πŸ‘‹ I'm Mona β€” I teach machines to make sense of messy language (and sometimes, I try to make sense of people).
  • πŸŽ“ Doctoral researcher in Machine Learning & Data Science, fortunate to be supervised by Professor Jaakko Peltonen.
  • 🌍 6+ years building machine learning systems for low-resource and multilingual NLP (and occasionally begging my GPU not to crash 🧊).
  • πŸ€– Into LLMs, representation learning, explainable AI, and making models play nicely across modalities without throwing a tantrum.
  • 🧠 My current work uses probabilistic ML to discover and track subtle and rare mental-health signals in youth online discourse (TUBEDU project)β€”responsible AI with a real human payoff.
  • πŸ”¬ I blend academic rigor with real-world engineering to make smart things smarter.
  • πŸ“¬ Let’s connect β€” always open to research collabs, postdoc opportunities, cool ideas, or just a good chat!

Research Experience

Doctoral Researcher

Tampere University β€” TUBEDU Project

Pursuing my Ph.D. in Computing Sciences with a focus on data science and statistical data analysis. My work centers on representation learning, topic modeling, and responsible AI for analyzing Finnish-language youth mental-health discussions on YouTube. I'm especially interested in uncovering hidden, under-represented structures in large, noisy datasets β€” minority-aware models that reduce harm from under-detected vulnerable content. Supervisor: Prof. Jaakko Peltonen.

Feb 2024 – Present

Research Assistant

Sharif Language Processing Lab (SLPL), Sharif University of Technology

Worked on NLP for low-resource and morphologically rich languages. My Master's research on formality style transfer in Persian combined linguistic theory with modern deep learning, alongside contributions to language generation and deep neural models. Supervisor: Dr. Hossein Sameti.

Aug 2021 – May 2022

Independent & Cross-Modal Research

NLP & Speech

In parallel with my formal research, I've worked on independent projects in speech processing and automatic speech recognition, together with cross-modal research connecting audio and text. These shaped my broader interest in multimodal learning and the joint modeling of signals across formats.

Job Experience

NLP Engineer & Researcher

Iran Telecommunication Research Center (ITRC)

Built benchmark datasets for Persian and researched Large Language Model (LLM) methods for low-resource settings β€” with a focus on prompt engineering, fine-tuning, and evaluation.

Jun 2023 – Dec 2023

Mentor β€” Speech & Language Processing

Hamrah Academy

Mentored students on speech and language processing fundamentals and practical pipelines, based on the course by Dr. Hossein Sameti, guiding learners through real-world audio-processing tasks.

Feb 2022 – Aug 2022

Researcher / NLP Engineer

ASR Gooyesh Pardaz Co.

Led development of real-world tools for Persian text and audio data, contributing to production speech and NLP systems:

  • Contributed to production ASR/TTS systems β€” Ariana (TTS) and Nevisa (ASR), including quantitative evaluation
  • Built Romand: a production-ready Persian Q&A assistant
  • Designed a Persian verb analyzer and a rule-based PDF text-extraction module for commercial use
  • Worked on speech recognition and text classification modules
Sep 2020 – May 2022

Quality Control Supervisor

ASR Gooyesh Pardaz Co.

Oversaw QA for annotation pipelines and speech-data alignment, ensuring accuracy and consistency in large-scale training corpora.

Sep 2020 – Aug 2021

Education

Tampere University

Doctor of Philosophy in Computing Sciences β€” Data Science & Statistical Data Analysis

Faculty of Information Technology and Communication Sciences (ITC)
Focus: Machine Learning, Data Science, and Mental-Health Analysis on online platforms
Supervisor: Prof. Jaakko Peltonen

Feb 2024 – Present

Sharif University of Technology

Master of Science in Natural Language Processing

CGPA: 18.02 / 20
Thesis: Formality Style Transfer Using Deep Neural Networks
Relevant courses: Machine Learning, Artificial Intelligence, Statistical NLP, Computational Linguistics, Syntax, Morphology, Semantics

Sep 2019 – Sep 2022

Publications

Author profiles: Google Scholar Β· ORCID

Peer-Reviewed
  • Ebrahimi, S. F. & Peltonen, J. (2025). Constrained Non-negative Matrix Factorization for Guided Topic Modeling of Minority Topics. EMNLP 2025, Suzhou, China. Association for Computational Linguistics. [ACL]
  • Ebrahimi, S. F. & Peltonen, J. (2026). Nonnegative Matrix Factorization for Joint Clustering and Topic Modeling with Minority Topics. ICONIP 2025. LNCS vol. 16309, Springer, Singapore. [DOI]
  • Ebrahimi, S. F. & Peltonen, J. (2026). Voices Between Lines: Interpretable Labeling of Mental Health Minority Topics with Seed Guidance and LLMs. ECML PKDD 2025. CCIS vol. 2842, Springer, Cham. [DOI]
  • Ebrahimi, S. F., Akhavan Azari, K., Iravani, A., Qazvini, A., Sadeghi, P., Taghavi, Z., & Sameti, H. (2024). Sharif-MGTD at SemEval-2024 Task 8. SemEval-2024. [ACL]
  • Ebrahimi, S. F., Akhavan Azari, K., Iravani, A., Alizadeh, H., Taghavi, Z., & Sameti, H. (2024). Sharif-STR at SemEval-2024 Task 1. SemEval-2024. [ACL]
  • Zahed, F., Ebrahimi, S. F., Bahrani, M., & Mansouri, A. (2023). Annotated Persian COVID-19 News for Fake News Detection. PACLIC-2023. [ACL]
  • Saghayan, M. H., Ebrahimi, S. F., & Bahrani, M. (2021). Impact of Machine Translation on Fake News Detection (Persian COVID-19 Tweets). ICEE-2021, 540–544. doi:10.1109/ICEE52715.2021.9544409.
Under Review
  • Ebrahimi, S. F. & Peltonen, J. Minority-Aware Trend Model: Tracking Weak Mental-Health Signal Evolution. Under review at Transactions on Machine Learning Research, 2026.

Research Funding & Grants

  • Tietotekniikan tutkimussÀÀtiΓΆ Grant (2026) β€” Research grant awarded by the Finnish Foundation for Information Technology Research to support Ph.D. research. [Foundation]
  • The Swedish Academy of Technical Sciences in Finland 2026 Grant (2026) β€” Competitive research visit to support Ph.D. research: Responsible Predictive Modelling of First Youth Mental-Health Self-Disclosure
  • TUBEDU Project Research Funding (Academy of Finland) (Feb 2024 – Jan 2026) β€” Ph.D. research funding for ML on youth mental-health analysis. [Project]
  • ITC Faculty Funding (2026) β€” Supplementary research funding to support Ph.D. studies.
  • Health Data Science Seed Funding (2026) β€” Awarded by the Faculty of Medicine.

Projects

  • Adaptive AI System for Continual Learning [Capstone, GPT Lab x Bittium, 2025] – Built a multimodal, continually updating AI using RAG, LoRA/QLoRA, and FastAPI. Integrated speech, vision, and text; deployed real-time with GitHub, ArXiv, and StackOverflow retrieval. [GitHub]
  • TUBEDU Project [Health Care, Mental Health] – Machine learning analysis of Finnish YouTube vlogs to uncover patterns in youth mental-health discourse. Focused on representation learning, topic modeling, and clustering. [Project]
  • Multilingual Machine-Generated Text Detection [LLMs, Evaluation] – Developed a black-box framework to detect machine-generated text across languages. Published in ACL@SemEval-2024.
  • Fact-Checking in the COVID-19 Era [Fake News, Dataset] – Built a Persian fact-checking corpus from COVID-19 news to support misinformation-detection research.
  • Sharif-STR & Sharif-MGTD @ ACL/SemEval [Benchmark, Transformers] – Participated in SemEval-2024 with transformer-based models for semantic relatedness and text authenticity.
  • Semantic Relatedness for Low-Resource Languages [Cross-lingual NLP] – Designed transformer models to evaluate semantic similarity in African and Asian languages.
  • Romand Intelligent Assistant [Dialog Systems] – Created and deployed a production-ready chatbot for real-world Q&A in Persian.
  • Farsquad [Style Transfer, Persian] – Developed a deep-learning pipeline and dataset for formality style transfer in Persian.

Technical Skills

Machine Learning & NLP

PyTorch, Hugging Face Transformers, TensorFlow, Keras, scikit-learn, Gensim, spaCy, NLTK, LLM fine-tuning, RAG, LoRA/QLoRA, PEFT

Programming & Scripting

Python, JavaScript, R, MATLAB, Java, Bash, SQL

Full-Stack Web

React, Node.js, Express, MongoDB, REST, JSON, React Router, Webpack, HTML5, CSS3, Bootstrap, FastAPI, Streamlit

Cloud & DevOps

AWS (EC2, S3, IAM, Auto Scaling, VPC), Docker, Linux, Git, GitHub Actions, CI/CD

Data Analysis & Visualization

Pandas, NumPy, Matplotlib, Seaborn, Power BI

Research & Reporting Tools

LaTeX, Jupyter, Overleaf, Microsoft Office

Languages

Kurdish (Native), Persian (Native), English (C1 β€” Academic/Professional), Finnish (A2 β€” Basic)

Awards & Recognitions

  • Merit Student Award, Sharif University of Technology (2021)
  • National Scholarship, Sharif University of Technology (2019)
  • Ranked 4th β€” National Graduate Entrance Exam (Iran) (2019)

Other Academic Merits

  • Invited Reviewer – NeurIPS 2026
  • Invited Reviewer – AISTATS 2025
  • Reviewer – TPDL 2025, Finland
  • Program Committee (PC) – TPDL 2026
  • Invited Reviewer – SemEval 2024, Mexico City
  • Reviewer – Natural Language Engineering journal, 2022
  • Technical Facilitator – ICWE 2024, Tampere (AV + session support)
  • Conference Volunteer & Attendee – EACL 2023, Dubrovnik, Croatia

Teaching

Teaching Assistant (Graduate Level)

Sharif University of Technology

Courses: Artificial Intelligence, Natural Language Processing, and Computational Linguistics. Supervisor: Dr. Mohammad Bahrani.

Oct 2020 – Sep 2021

Certifications

  • Full Stack Open β€” Deep Dive into Modern Web Development, University of Helsinki: React, Node.js, MongoDB, REST, GraphQL, CI/CD, Docker
  • Microsoft Power BI Full Course β€” Power Query, DAX, dashboards, multi-source data modeling

Interests

Representation Learning – Learning structured patterns from data across modalities, especially in language and speech.

Social Media & Mental-Health Analysis – Investigating behavioral trends and narratives from large-scale online discourse.

Multimodal Machine Learning – Integrating text, audio, and visual data for richer modeling and understanding.

Speech & Language Technologies – Building systems that understand and generate human language, with a focus on low-resource contexts.

Interpretable & Constrained Models – Designing AI with structure, transparency, and human alignment in mind.

Applied Machine Learning – Transforming research ideas into real-world tools for meaningful impact.