Germany
I'm a hands-on ML/AI Engineer with a strong foundation in applied mathematics and deep experience across industries like manufacturing, e-commerce, healthcare, and online retail. I specialize in building scalable, production-ready AI systems — from forecasting models and custom NLP pipelines to intelligent chatbots with multimodal capabilities. 🔍 What I Do: Design and deploy AI solutions using cutting-edge techniques such as self-trained YOLO object detection, transformer-based time-series forecasting (Autoformer, LSTM, GRU), and function-enhanced LLMs. Build multilingual chatbots using ChatGPT with custom RAG + CAG architectures for high-performance question-answering and voice interaction. Integrate OCR, speech-to-text, and web crawling into intelligent systems to enhance data accessibility and interactivity. Create fast, scalable backend services with async FastAPI and optimize workflows using MLFlow, PyTorch Lightning, and custom tokenization. 💬 Leadership & Communication: I’ve led teams, managed AI project pipelines, and acted as a communication bridge between technical teams and stakeholders. 🎓 Academic Background: With a Master's in Applied Mathematics for Network and Data Science (Germany) and a Bachelor's in Mathematics (Turkey), I bring strong analytical thinking to every AI challenge I tackle. Always open to connecting with professionals passionate about AI, data science, and building intelligent systems that make a difference.
Leading enterprise AI initiatives at Vogelsang, architecting production-grade AI systems that bridge external LLM capabilities with internal infrastructure. AI Infrastructure & LLM Systems - Architected enterprise MCP server integrating external LLMs with internal knowledge bases and tooling - Deployed secure, on-premise agentic chatbot with bad prompt detection, anonymization, and token optimization - Conducted red team security assessments for prompt injection and adversarial attack remediation - Built custom AI agent tools: geolocation, email automation, helpdesk ticketing, web search, intelligent conversation routing Advanced RAG & Knowledge Systems - Developed production RAG system on PostgreSQL with pgvector for high-accuracy semantic search - Engineered agentic RAG pipeline with group-level access controls and AI-powered context extraction - Built automated data (html, pdf, image) pipelines with daily updates and/or event-driven triggers, and multimodal CI/CD ingestion Model Development & Optimization - Fine-tuned Hugging Face models using PEFT/LoRA and optimized OpenAI models for domain tasks - Implemented time-series forecasting with Gradient Boosting, GRU, LSTM, and Autoformer architectures - Created custom tokenizer for product-specific patterns, improving LLM accuracy - Fine-tuned TAPAS for German-language tabular question answering Backend & Data Engineering - Built scalable FastAPI backend services for AI model hosting and microservices - Applied NER for automated sensitive data detection and anonymization - Managed PostgreSQL databases and deployed Metabase for real-time AI monitoring - Integrated speech-to-text systems for voice-enabled interactions Leadership - Primary technical contact for AI initiatives and stakeholder communication - Mentor interns and coordinate cross-functional workflows
- Forecasted monthly revenue using SARIMAX models to support strategic financial planning. - Built and maintained interactive dashboards in Tableau to visualize key performance indicators and business trends. - Designed and managed robust ETL pipelines using Airflow, Python, and BigQuery for scalable and efficient data workflows. - Created, validated, and optimized data sources to ensure data quality and reliability for analytics. - Acted as a liaison between technical teams and stakeholders, translating business needs into actionable data solutions and ensuring alignment across teams.
- Worked on healthcare data - Processed image data Image to use with machine learning techniques - Analyzed the patterns of IFE blood test data with Python
- Defined and implemented CSV data validation rules using JSON Schema to ensure data integrity and consistency in customer datasets. - Supported the development and packaging of local Python modules using PySpark and Poetry to streamline code reuse and collaboration. - Performed large-scale data analysis using PySpark on Databricks, enabling efficient processing and insight generation from big data. - Gained hands-on experience with Confluent Kafka by assisting with message streaming and simple data pipeline tasks in a distributed system.
- Processed and structured large volumes of image data into NumPy objects to create machine learning-ready datasets. - Performed pattern recognition on Immunofixation Electrophoresis (IFE) blood test data to identify anomalies and assist in medical data interpretation. - Analyzed outcomes to extract meaningful insights and support further research and model development.