Emre AYDIN

Lead a 10-person data team, structured into two verticals:

• AI Applications – focused on machine learning systems and computer vision solutions
• Reporting & BI – responsible for data infrastructure and company-wide analytics

• Built a full-stack Dynamic Pricing system powered by demand forecasting, driver supply, and price elasticity models, enabling real-time fare and payout optimization.
• Developed an OCR-based license plate verification service, automating vehicle data validation and improving operational efficiency by 35%.
• Delivered a facial recognition system for secure user verification, reducing manual review efforts by 60%.
• Collaborated with cross-functional teams to ensure ML systems were production-ready, interpretable, and aligned with business goals.
• Conducted solution architecture reviews and vendor evaluations to align external capabilities with internal modeling strategies.
• Modernized data pipelines using AWS (Athena, Glue, S3) and orchestrated workflows via Airflow and Flink for scalable real-time analytics.
• Delivered business-critical dashboards to operations, finance, and growth teams, enabling faster and more informed decisions.
• Drove cross-functional collaboration among analysts, engineers, and scientists to produce cohesive, end-to-end data products.

Tech Stack

Python, AWS (Athena, Glue, S3, EKS), Apache Airflow, Apache Flink, PostgreSQL, Docker, Kubernetes, OpenCV, TensorFlow, PyTorch, scikit-learn, Pandas, Git, Jupyter

• Led a 9-person data team, delivering tailored AI and analytics solutions for clients across automotive, retail, and infrastructure sectors.
• Drove cross-industry applications of machine learning with a focus on campaign optimization, predictive modeling, and customer segmentation.
• Conducted Campaign ROI Analysis for Arçelik, enabling data-driven marketing decisions and measurable returns.
• Led Campaign Optimization for Tofaş, increasing promotional efficiency and targeting accuracy through data-driven simulations.
• Developed a Next Vehicle Purchase Prediction Model to improve demand planning and inventory control for an automotive client.
• Performed Value Segment Analysis for Opet, identifying high-value segment migrations for more effective customer lifecycle strategies.
• Delivered Traffic & Campaign Analytics for the Eurasia Tunnel, enhancing user experience and operational planning through granular insights.
• Initiated infrastructure renewal projects to modernize internal data operations and improve model deployment efficiency.
• Organized and delivered internal training sessions to upskill team members and ensure adoption of latest data science techniques.

• Conducted large-scale web data extraction from Crunchbase, building robust crawlers capable of bypassing bot detection mechanisms.
• Explored the use of Large Language Models (LLMs) for generating web content summaries, leveraging vector databases for efficient semantic search and retrieval.
• Delivered prototype systems integrating OpenAI APIs with Pinecone to test scalable document understanding workflows.

Tech Stack:

Python, Selenium, BeautifulSoup, Requests, OpenAI API, Pinecone, Vector Databases, Jupyter, Git

• Led end-to-end development of scalable AI solutions across telecom, HR, and digital services, with a focus on automation, personalization, and customer interaction.
• AutoML Platform:
  Co-led the design and launch of Turkcell’s internal AutoML platform, enabling non-technical users to train models on their own datasets. Generated initial revenue of €45,000 from B2B commercialization.
• Smart SMS Categorization:
  Developed a natural language processing system to auto-correct and classify SMS content, contributing to ₺12.5M in cost savings through automation.
• AI-Based Video Interview Scoring:
  Delivered AI models to evaluate over 60,000 candidate applications for Turkcell’s talent programs by analyzing resumes and interview videos using computer vision and NLP techniques.
• TBOT Chatbot Platform:
  Built the NLU (Natural Language Understanding) module and deployed it on OpenShift as part of Turkcell’s enterprise chatbot system. Used across internal services and external municipalities.
• Played a key role in cross-functional AI governance and internal community building, promoting best practices and knowledge sharing across teams.
• Actively contributed to event organization and developer training, fostering engagement with cutting-edge AI tools and techniques.

Tech Stack

Python, TensorFlow, PyTorch, Keras, OpenCV, Scikit-learn, Docker, OpenShift, FastAPI, NLP, Text Classification, Video Processing, Git, Jupyt

• Delivered scalable data science and engineering solutions across personalization, campaign targeting, and behavioral analytics for one of Turkey’s largest retail groups.
• Clickstream Data Platform:
  Built a multi-cloud data ingestion system capable of processing 90M+ events per day, supporting real-time customer behavior tracking and analytics.
• Personalized Product Ranking (Smart Listing):
  Designed and deployed ML models to score all products for all users on beymen.com, improving:
  → Add-to-basket conversion by +31.8%
  → Category-to-product click-through by +10.5%
• Campaign Optimization Engine:
  Developed targeting models to match users with the most relevant campaigns (via Hopi), outperforming legacy methods in 5 of 6 test campaigns.
• Integrated data science outputs into CRM and marketing platforms, enabling automated personalization and segmentation.
• Collaborated across marketing, product, and engineering to ensure alignment of modeling efforts with business impact.
• Took ownership of end-to-end R&D: from problem framing to deployment in production systems.

Tech Stack

Python, Scikit-learn, Clustering, PostgreSQL, Oracle SQL, Apache Airflow, Google Cloud Platform, AWS, Docker, Kubernetes, Hive, BigQuery

• Contributed to the development of a customer service chatbot, implementing backend services, chatbot infrastructure, and a custom UI for internal use.
• Applied text similarity and text classification techniques using machine learning to improve chatbot response accuracy in Turkish.
• Built a machine learning model to recommend solutions for customer complaints, using text vectorization and NLP-based classification.
• Developed a prototype system that captures images at intervals and uploads to Instagram using Python and OpenCV.

Tech Stack

Java, Spring Boot, Weka, D4LJ, AngularJS, Python, OpenCV, NLP, Text Classification, Instagram API, Git