Istanbul, Türkiye
At Mirsis Bilgi Teknolojileri, contributing as a Data Scientist, my work focuses on integrating artificial intelligence to streamline operations and enhance business outcomes. With over two years of experience in data science and AI, I specialize in developing automated solutions and uncovering actionable insights through advanced analytics, data visualization, and machine learning tools such as LLMs, Scikit-Learn, and Python. Recent achievements include transforming legacy systems into smart, automated workflows and driving innovation through collaborative research initiatives. Proficient in leveraging deep learning and NLP techniques, my contributions empower teams to make data-driven decisions and deliver impactful results. Dedicated to advancing the possibilities of AI, I aim to continue creating systems that redefine operational efficiency and innovation.
• Analyze complex datasets to uncover insights that drive business growth through data visualization, analytics, and AI tools • Contribute to research initiatives and develop innovative solutions to enhance company operations and services • Optimize existing systems by transforming legacy applications into smart, automated processes Technical Skills: Data Visualization Tools (Tableau, Matplotlib, Seaborn, Plotly), Python, SQL, Scikit-Learn, LLM
• Sales Prediction: Reduced RMSE score by 8% in ongoing project. • Recommender System: Developed a reusable framework for market basket analysis, reducing development time and increasing sales by 5%. • Customer Segmentation: Created a predictive system for new customer sales with a 12% RMSE error reduction. • Code Translation: Converted existing COBOL codebase to JAVA using a large language model (LLM). • Customer Support Segmentation: Designed a system for routing user complaints using LLM. Keywords: Statistics, Scikit-Learn, Deep Learning (TensorFlow, MLP/DNN, LSTM, GRU), NLP (Transformers, Embedding), LLMs (GPT-3.5, LLaMa model fine-tuning), MS Azure, SQL (MySQL, Hadoop/Hive, SQLAlchemy, PyODBC), Spark, Polars, Data Visualization (Matplotlib, Seaborn), Seasonal modelling (ARIMA, SARIMA), Explainable AI (SHAP, Shapash), Imbalanced Learning, Data Formats (CSV, JSON, XML, Parquet).
• Speech-Lip Synchronization: Developed a system for realistic animation character speech employing both computer vision and signal processing techniques. • Written Exam Auto-Grader: Created an auto-grader for handwritten exam papers with less than ±7 points deviation. Keywords: Deep Learning (PyTorch, LSTM, GRU, CNN, ANN, Siamese Network, Transformers, GAN), Linux, Computer Vision (OpenCV, DeepFace, PyImage), NLP (Transformers, Embedding), Signal/Voice Processing (Librosa, PyAudio, Soundfile), OOP, Data/Image Augmentation, Web Scraping (BeautifulSoup4, Scrapy, ffmpeg), Git, Docker, Parallel Processing (Thread, Multiprocess, Joblib), Pipelining, Data Generators, ArcGIS, Erdas Imagine, Geospatial Data Processing, Raster Data Processing.
• Mentoring: Guided trainees to achieve their goals with high satisfaction rates. • Content Writing: Developed new content and instructional materials to enhance learning. • Datathon Jury Member: Evaluated participants in a recruitment event for a prominent Turkish company. Keywords: Python, Machine Learning (Supervised, Semi-Supervised, Unsupervised), Deep Learning, Mathematics, Data Visualization (Matplotlib, Seaborn), Relational Databases, Databricks, GitHub, Numpy, Pandas, Statistics, Bash Programming.
• Demonstrated accountability by evaluating military / geographical data on raster maps using ArcGIS to extract valuable insights, • Managed complexity and risk by exploring 50+ new formations and reporting to authorities, awarded for outstanding success, • Founded new sources for collecting military data and a new validation method through multiple combinations,
Responsible for training and management of a team of up to hundred people.