Student data scientist working on applied ML and LLM projects for healthcare and industry partners.
- OffRa Health: Built an ML and LLM system to predict breast cancer risk and generate personalized diagnostic care plans;
implemented a novel research method on medical data; deployed a live web app under clinician review.
- Schematics: Designed a three-stage ML pipeline (two classifiers and regression) to replace bid-based ranking with expected-profit optimization;
built and evaluated models in Databricks; delivered an internal decision-support tool.
- Assuta Ashdod Hospital: Developing an NLP system to predict fasting requirements from medical records using locally hosted LLMs with Ollama.
Researcher
Ben-Gurion University of the Negev | Israel’s Directorate of Defense Research and Development (MAFAT) | 2024-Present
Collaborating with Israel’s Directorate of Defense Research and Development (MAFAT) to assess how well large language models (LLMs)
preserve an article’s stance in the summaries they generate, with a focus on Hebrew.
This research aims to enhance the accuracy and cultural relevance of LLMs in Hebrew, addressing the unique linguistic challenges associated with this language.
My work involves metric design, data analysis, and testing on state-of-the-art models.
Teaching Assistant for Databases
Ariel University | Apr 2024 - Aug 2024
Ben-Gurion University of the Negev | Oct 2025- Present
My responsibilities include instructing students in MySQL, Relational Algebra, MongoDB, and Neo4j,
as well as statistical techniques such as Naive Bayes, logistic regression, and linear regression.
Prompt Engineer
Code for Israel | Knafayim Shel Krembo | AI Team - Krembot Group | 2025-Present
Developing AI-driven educational games to enhance learning experiences for members
with special needs, focusing on accessibility and engagement.
Research Assistant
Ariel University | Oct 2021 - Sep 2024
2023 - 2024:
Engaged in research and development of a Hebrew Large Language Model (LLM) derived from llama2,
employing PyCharm IDE alongside PyTorch and Transformers libraries.
2021 - 2022:
Published article in Scientific Reports of Nature Journal on
"Classifying interpersonal synchronization states using a data-driven approach:
implications for social interaction understanding".
Utilized machine learning and data exploration techniques,
employing Python in Jupyter Notebook for analysis and model development.