I collaborate closely with Schneider's Global Marketing team as a Data Scientist to glean actionable insights from data. I use Python for sophisticated data interpretation, improving and organizing datasets to help with our marketing plans. One of my duties is using Tableau to create dynamic visualizations that help the team make data-driven decisions. In order to make sure that our marketing initiatives are supported by solid data insights, I also mentor new team members, assisting them in honing their data analysis and visualization abilities.

I work at UDENE where I help create predictive models and workflows for road data and land surface temperature data, utilizing GIS technology to improve urban planning. My job involves using React and Django to build and improve a full-stack exploration tool. I also integrate machine learning solutions to offer deeper insights. With the use of Earth Observation data, this tool helps urban planners make evidence-based decisions by simulating and analyzing various urban development scenarios. My work is essential to bringing the project's goals into line with the Sustainable Development Goals of the UN, especially when it comes to strengthening the resilience and sustainability of cities.

Working closely with industry partners and TÜBİTAK to develop and implement innovative solutions in data science and artificial intelligence.In my current role, I am deeply involved in a project focused on detecting Hepatocellular Carcinoma (HCC) using patient data. This work has included the development of sophisticated models utilizing DCGANs (Deep Convolutional Generative Adversarial Networks) to enhance our dataset, allowing us to generate synthetic data to augment the original patient data. To ensure the accuracy and reliability of these models, we implemented confidence thresholding, which further refined the data by filtering out less reliable generated images. Building on this enhanced dataset, we developed and trained EfficientNet models, which have shown significant improvements in predictive performance for HCC detection.In addition to my work on HCC detection, I have also contributed to the creation of news classification models. These models are designed to categorize news articles based on their content, leveraging advanced NLP techniques to ensure accurate and meaningful classifications. This project has broad implications for the field of natural language processing and has practical applications in media and information dissemination.My role goes beyond mere development; it involves rigorous academic research and the publication of findings in scholarly papers. This position allows me to apply my data science and AI expertise to solve real-world problems while staying at the cutting edge of both academic research and industrial applications.

Company Analytics with NLP: Utilized BERT, Word2Vec, TF-IDF, and MinHash LSH for advanced company vector analysis.Medical Image Generation: Innovated with GAN and DCGAN models in healthcare imaging.Predictive Web App Development: Created a contract prediction tool using Next.js and FastAPI, blending data science with full-stack web development.

Selected as one of 75 students selected from Turkey's most prestigious universities in engineering.6 months-long program specializing in Data Science, ML, Al, Microsoft solutions and project development.

As part of an intern team at Jotform, I fully developed a full-stack fraud detection application using Next.js for the frontend and FastAPI for the backend. The app addressed a variety of fraud use cases, including account stealing and credit card theft, by integrating BERT for classification and BERTopic for categorizing fraudulent activities. I also contributed extensively to training the models on form data to detect suspicious patterns. Our application, which integrated with Jotform's API, allowed users to analyze both their own forms and third-party forms for potential fraud. This experience honed my skills in full-stack development, machine learning, and fraud detection.