FURKAN DEMIR

APPLIED AI & DATA ANALYST

Bridging statistics, automation, and generative AI to solve complex business problems.

MSc Statistics & Data Science @ KU Leuven

Furkan Demir

THE MISSION

I build intelligent systems that turn complex data into actionable strategy. With a background in Statistics and a focus on Applied AI, I specialize in automating expert-level workflows—from multitenant benchmarking to linguistic compliance monitoring.

My work currently focuses on the intersection of Finance and AI, helping organizations in the Belgian banking sector navigate digital transformation through RAG architectures and custom NLP pipelines.

EXPERIENCE

SAILPEAK

2025.03 - PRESENT
Data Analyst (AI & Automation, Part-Time)

Brussels, Belgium

  • Contributed to the development of a benchmarking dashboard covering 44+ European retail banks, focusing on GEO, Product Acquisition and Experience. Led the full data collection and handling process.
  • Built the design and automation of a CEFR B2 compliance assessment system, leveraging Gemini AI to evaluate web content under the EU Accessibility Act.
  • Worked on the design of a retrieval-augmented (RAG) assistant for Belgian banking products, together with colleagues, focusing on accounts, cards, and fee structures.

VODAFONE

2024.02 - 2024.09
Big Data & AI Discover

Istanbul, Turkey

  • Built binary classification models (XGBoost) for high-accuracy customer behavior prediction.
  • Developed an automated GenAI newsletter pipeline on Azure to streamline internal tech communication.
  • Awarded "Vodafone Stars" for taking part in company-wide prompt engineering initiatives.

VODAFONE

2023.08 - 2024.02
Customer Analytics Discover

Istanbul, Turkey

  • Utilized SQL and advanced Excel for large-scale customer call data analysis.
  • Optimized customer experience and business performance through data-driven insights.

GALATAPORT ISTANBUL

2022.08 - 2023.01
Data Analytics Intern

Istanbul, Turkey

  • Built operational data reports for senior management to optimize project delivery.
  • Managed digital content lifecycle for mobile platforms, ensuring data accuracy across user touchpoints.

EDUCATION

KU LEUVEN

2024.09 - Present
MSc in Statistics and Data Science

Leuven, Belgium

BOĞAZİÇİ UNIVERSITY

2018 - 2024
B.A. in Economics

Istanbul, Turkey

SELECTED WORK

Retail Banking Website Benchmark 2025

AI-powered evaluation of 44 leading banks across 6 EU countries (BE, FR, NL, DE, IT, ES).

  • Scale: Analyzed 10,000+ data points across Discoverability, Acquisition, and Experience categories.
  • GEO Focus: Designed metrics for Generative Engine Optimization to assess how banks appear in AI search engines.
  • Strategy: Delivered data-driven C-level insights on WCAG 2.2 and CEFR B2 accessibility compliance.

Belgian Finance RAG (Ongoing Project)

AI-powered assistant specialized in Belgian banking products and fees.

  • Scope: Natural language interface for 9 major Belgian banks (accounts, cards, and services).
  • Tech: Full-stack RAG application grounded in real banking data extracted via custom crawlers.
  • Value: Simplifies complex fee structures and product comparisons for end-users.

Topic Modeling and Thematic Analysis of JFK Assassination Files Using NLP (Master Thesis)

End-to-end NLP pipeline for 83,621 declassified pages. [Ongoing]

  • Scale: Architected a multi-stage pipeline to index, categorize, and extract text from 2,566 documents (83,621 pages).
  • OCR Engineering: Benchmarked 8 state-of-the-art methods (GCP Vision, DeepSeek-OCR, Tesseract) to achieve optimal accuracy.
  • LLM Intelligence: Automated document categorization using Groq Llama models and Leo Prompt Optimizer.
  • Analysis: Applying BERTopic and LDA to detect historical pattern shifts and thematic evolution across archives.

Zanaat.app ↗

Full-stack interactive craft game exploring traditional artisanal skills.

  • → Designed gameplay mechanics and craft-matching logic for an educational experience.
  • → Implemented real-time UI/UX to engage users with traditional crafts.

CEFR B2 Compliance Engine

Automated linguistic auditing system for EU financial accessibility standards.

  • NLP scoring: Evaluates vocabulary, grammar, and complexity using structured LLM prompts.
  • Automation: Built crawling and evaluation pipelines for multilingual banking pages.

Operations & Budget Automation

End-to-end automation of financial operations and employee benefits.

  • Mobility & Learning Budget: Built a system to track and manage employee cross-benefit programs via n8n and Google APIs.
  • Invoicing: Fully automated timesheet processing, PDF generation, and Gmail delivery using Apps Script.

CORE TOOLSTACK

🧠 AI, ML & NLP

LLMs (OpenAI, Gemini) Prompt Eng (XML/JSON/Eval) RAG (Embeddings/Semantic Search) Topic Modeling (LDA/SOM) OCR (GCP Vision, Tesseract)

📊 Data Science

Python, R, SQL Pandas & Scikit-learn Statistical Modeling Regression & Clustering

🤖 Automation

Selenium & BS4 Google Apps Script n8n & Zapier Gmail/Drive APIs

🗄️ Storage & Search

SQLite (FTS5) Semantic Search Hybrid Indexing

🌐 Web & Compliance

Sitemap Crawling CEFR B2 Scoring WCAG 2.2 GEO Metrics

📉 Viz & Cloud

Matplotlib & ggplot Google Cloud Platform Microsoft Azure Git & GitHub

LET'S CONNECT

Always open to discussing applied AI, data strategy, or research collaboration.