Gülten MERİÇ

Developed and implemented advanced reverse engineering and data scraping solutions for collecting and processing business-critical data from web-based systems.

Designed custom algorithms and strategies to bypass multi-layered anti-bot systems and CAPTCHA mechanisms, ensuring 80% uninterrupted data flow.

Reverse engineered hidden API endpoints in dynamic JavaScript applications, accelerating data integration processes by approximately 75%.

Developed parsers to extract structured data from websites with complex and constantly changing DOM structures with millimeter precision.

Optimized existing data collection processes by deeply analyzing browser traffic (HTTP/S, WebSocket).

Ensured data availability by developing proactive solutions (smart proxy rotation, browser fingerprint management) against IP blocking and rate-limiting issues.

Advanced Web Reverse Engineering:
JavaScript Dynamic Analysis & De-obfuscation, Hidden API Discovery & Exploitation, Anti-Bot & CAPTCHA Bypass Strategies.

Web Scraping & Data Extraction:
Fine-grained Data Parsing from Complex & Dynamic DOM Structures (XPath, CSS Selectors).
Deep Crawling with Headless Browser Automation.
Request-based High-Performance Scraping.

Programming Languages & Libraries:
Python with Scrapy, BeautifulSoup, Requests, Playwright/Selenium integration.

Tools & Environments:
Browser Developer Tools; Chrome DevTools, Firefox Developer Tools, Proxies, VPNs, Git, Docker.

AI-Powered Assistant System for Employee Support (End-to-End, Production Deployment)

A complete AI-driven assistant solution built for Çilek Mobilya, combining advanced document retrieval, chatbot response generation, and mobile user interaction.

Significantly improving internal communication and efficiency. Currently utilized by over 100 employees daily, streamlining access to critical information.

Backend: Developed in Python (Flask) with a full RAG (Retrieval-Augmented Generation), including text embedding architecture integrating Gemini 2.0 Flash, OpenAI GPT-3.5, Mistral, and FAISS Vector Database.

Utilized BERT-based embedding models for semantic similarity search, allowing highly relevant document chunks to be retrieved and used as context for LLM responses.

Extracted context from internal PDFs using PyMuPDF and Tesseract OCR; Watchdog enabled real-time document monitoring.

Frontend: Built a Flutter-based mobile interface for employees to check shifts, submit leave requests, and chat with the assistant.

Deployment: System fully integrated via RESTful APIs and deployed in multiple production environments including; Linux server using Webmin, Windows Server with IIS, Cloud environment using Azure.

• Worked on a WhatsApp chatbot project using Python, Dialogflow, and Twilio API.
• Performed data preprocessing and feature engineering for various AI models, ensuring data quality and model performance.
• Applied natural language processing (NLP) techniques to develop and refine text-based AI solutions.
• Contributed to the development and evaluation of machine learning algorithms throughout the project lifecycle.

• Developed a virtual market application using C# and .NET, and SQL with databases (MSSQL) enabling users to list, filter, add products to the cart, and place orders.
• Optimized the application for a user-friendly interface and fast processing performance.

• Collaborated with the Google DSC Teknofest team to develop a mobile application project in the 'Barrier-Free Life for People with Disabilities' category.
• Enhanced teamwork, problem-solving, and technical design skills during the project.
• Participated in the Solution Challenge competition with our project.
• Worked well with teammates and accepted coaching from management team.