Currently serving as a Teaching Assistant for DSA 210, I support students in mastering the foundational pillars of data science, including computational thinking, machine learning, and statistical inference.

I lead weekly office hours to provide hands-on guidance with real-world data tools and techniques, helping students navigate complex topics such as exploratory data analysis (EDA), data privacy (K-anonymity), and predictive modeling.

By bridging the gap between theoretical coursework and practical application, I assist my peers in developing the technical rigor necessary to solve diverse problems across healthcare, business, and social domains.

During the Fall 2025-26 term, I conducted a comprehensive scientometric study analyzing over 250 global publications to map the evolving landscape of AI in healthcare.

My research focused on the critical intersection of explainable AI (XAI), clinician trust, and regulatory governance.

By utilizing literature mining and network analysis, I identified key knowledge gaps and emerging trends necessary for developing trustworthy and clinically deployable AI systems.

The LinkedIn Post

As the Project Lead for DeepMyelinAI, I head a multidisciplinary team developing an explainable AI-based clinical decision support system for Multiple Sclerosis (MS).

I oversee the integration of multimodal data—including CSF biomarkers, genetic risk profiles, and longitudinal clinical disability scores—into a unified machine learning framework designed for early diagnosis and disability progression prediction.

My role involves managing the full development lifecycle: from engineering Temporal Transformer architectures for longitudinal tracking to implementing XAI layers (SHAP/LIME) to ensure clinical transparency.

By aligning our system with the 2024 McDonald Criteria, I ensure our technical solutions meet rigorous medical standards, ultimately aiming to provide neurologists with actionable, data-driven insights.

Developing AI models to reduce unnecessary sepsis testing in Emergency Room settings.

Focusing on clinical efficiency by optimizing diagnostic triggers to reduce physician workload and healthcare costs.

Work In Progress...

In this project, I developed a high-precision machine learning model to predict the pathogenicity of variants in the COL1A1 and COL1A2 genes, which are primary causes of Osteogenesis Imperfecta (OI) — a rare genetic disorder affecting bone fragility.

Using a curated dataset from ClinVar, I engineered biologically meaningful features and trained multiple classifiers, ultimately achieving performance that surpassed established tools such as SIFT and general-purpose diagnostic predictors.

The pipeline included variant filtering, feature selection, cross-validation, and interpretability analysis — resulting in a reproducible and clinically relevant bioinformatics workflow.

View on GitHub →

Under the supervision of İleyna Sevde Üvak, I contributed to a molecular biology research project focused on computationally optimizing PETase enzymes for efficient degradation of PET plastics.

I applied computational biology techniques — including protein structure analysis and mutation screening — to identify candidate enzyme variants with improved thermostability and catalytic efficiency.

This work sits at the intersection of synthetic biology and environmental sustainability, exploring how AI-driven enzyme engineering can accelerate the development of biological solutions to plastic pollution.

As a Board Member and the Head of the BioTech Division at Sabancı University's Deep Technologies Club, I lead initiatives that bridge cutting-edge biology with emerging technology.

My role involves organizing workshops, seminars, and hands-on projects that expose students to topics like CRISPR, synthetic biology, computational genomics, and AI-driven drug discovery.

I also coordinate interdisciplinary collaborations between CS, biology, and engineering students — cultivating a community of innovators passionate about the future of life sciences and deep tech.