Sina Amiri Golilarz

• Collecting, cleaning, and preprocessing data from diverse sources such as databases, APIs, files, and sensors
• Standardizing datasets by handling missing values, inconsistencies, and errors to prepare data for analysis
• Performing exploratory data analysis (EDA) to identify patterns, correlations, and data distributions
• Applying statistical techniques to extract actionable insights from raw data
• Creating effective data visualizations using tools such as Python (Matplotlib, Seaborn) and Power BI
• Building predictive and machine learning models including regression, classification, and clustering
• Evaluating model performance using metrics such as Accuracy, F1-score, ROC-AUC, and improving models for real-world deployment
• Interpreting analytical results and translating complex outputs into clear business insights for stakeholders
• Developing interactive dashboards and analytical reports to support data-driven decision-making
• Developed and optimized machine learning models using XGBoost and Scikit-learn for classification and predictive analytics tasks.
• Utilized Dask for parallel data processing and handling large-scale datasets, improving data pipeline efficiency and reducing computation time.
• Collaborating with cross-functional teams (Product, Marketing, HR, Finance) to define data-driven problems and solutions
• Translating business requirements into data-oriented analytical tasks
• Supporting teams through data literacy training and domain-specific guidance
• Monitoring deployed models, updating them with new data, and maintaining data pipelines
• Documenting workflows, analytical processes, and codebases to ensure scalable future development
• Developing and advancing a high-performance Persian OCR system as a core product, utilizing OpenCV (cv2), YOLOv5 for text detection, PaddleOCR for recognition, and associated libraries
• Implementing end-to-end Persian document OCR pipeline, including image preprocessing, text localization with YOLOv5, character recognition and post-processing for right-to-left Persian script
• Achieving significant progress in accurate extraction of Persian text from diverse scanned/invoiced documents, leading to a production-ready product with strong performance in real-world scenarios
• Fine-tuning models and optimizing the OCR pipeline for improved accuracy, speed, and robustness on Persian-specific challenges (ligatures, diacritics, varied fonts and layouts)

• Conducted workshops introducing AI concepts and applications.
• Designed interactive exercises to enhance participant engagement.
• Supported students in completing practical AI tasks.

• Successfully implemented academic and industrial AI projects.
• Collaborated with multidisciplinary teams to solve complex problems.
• Applied state-of-the-art machine learning and deep learning techniques.
• Ensured performance optimization and task-specific tuning of models.
• Monitored project outcomes and reported results to stakeholders.
• Designed and implemented industrial and academic AI projects, including RAG-based knowledge
systems.
• Built vector databases using FAISS and ChromaDB for semantic search.
• Developed Agentic AI solutions combining reasoning and retrieval.
• Integrated LLMs (GPT, LLaMA, DeepSeek) into intelligent automation workflows.
• Applied advanced optimization and fine-tuning techniques to improve performance.

• Delivered lectures on advanced Python programming to undergraduate students.
• Developed and updated course materials and practical exercises.
• Guided students in hands-on programming assignments and projects.
• Assessed student performance and provided constructive feedback.

• Taught Python programming and AI fundamentals to diverse learners.
• Implemented real-world exercises for skill development.
• Mentored students on AI project implementation.