Hakan Gecili | Senior Data Scientist
Expert in Machine Learning, Optimization, and Computer Vision
About Me
I’m a senior data scientist and operations research analyst with over 10 years of experience designing and implementing machine learning and optimization solutions across industries like retail, logistics, and fintech. I hold a Ph.D. in Industrial Engineering and specialize in turning messy, high-dimensional data into intelligent systems that drive real business impact.From detecting fraud using graph-based ML at TransUnion, to optimizing shelf layouts for retail giants, to implementing computer vision systems that monitor store inventory in real-time — my work combines deep technical knowledge with business intuition.
💡 What I can help you with
Machine Learning & Predictive Modeling
Optimization & Operations Research
Computer Vision Solutions
Analytics Strategy & Prototyping
Model Governance & Automation
🛠 Skills & Tools
Programming & Scripting
Python • R • SQL • Java • C++ • MATLABMachine Learning & Hyperparameter Tuning
Scikit-learn • XGBoost • LightGBM • PyTorch • OptunaOptimization & Operations Research
CPLEX • Gurobi • AIMMS • MILP • Heuristics • Constraint ProgrammingComputer Vision
OpenCV • CNNs • Image Preprocessing • Anomaly Detection • Feature EngineeringData Handling & Visualization
Pandas • NumPy • Snowflake • Matplotlib • SeabornBig Data & Platforms
Apache Spark • AWS • GCPDocumentation & Reporting
Jupyter • docxtpl • Automated Model Governance
🚀 Selected Project
Device Recognition for Fraud Detection
Client: TransUnion
Tech: XGBoost, LightGBM, Neural Networks, Graph NetworksLed a team to build a high-performance device recognition system combining ML and graph-based identity resolution. Achieved 99.9% AUC-ROC and AUC-PR. Integrated advanced sampling and similarity strategies to enhance prediction on web, iPhone, and Android devices.
Small Aircraft Network Optimization
Client: UPS
Tech: CPLEX, GUROBI, PythonDesigned a strategic MILP model for long-term air network scheduling. Matched aircraft types to destinations and optimized flight paths and schedules. Built supporting cost models using regression.
Shelf Space Optimization for Retail StoresClient: Kroger Co.
Tech: Optimization (MIP), R, PythonDeveloped joint shelf design and space allocation models across 109 stores. Achieved up to 22% profit improvement. Created tools to guide assortment planning, layout optimization, and restocking strategies.
Computer Vision for Retail Inventory Monitoring
Client: Kroger Co.
Tech: OpenCV, CNN, PythonImplemented a CV pipeline to track product presence on shelves. Achieved 98% accuracy on select SKUs. Enabled real-time demand monitoring and reduced stockouts without manual audits.
Featured Publication
Joint shelf design and shelf space allocation problem for retailers
🌱 Personal Side
Outside of data science, I’m a lifelong learner, a problem-solver at heart, and someone who loves to improve systems — from machine learning models to home repair workflows. My background in optimization doesn’t just shape my professional work; it influences how I approach daily life.I’m an avid tennis and volleyball player, and I organize games and equipment pools for my community across four active groups. I commute by motorcycle, enjoy woodworking and gardening, and love fixing up bikes with my son. I have two cats and a history with dogs, and I enjoy helping friends with small repairs and DIY projects.I served 12 years in the Turkish Navy as an Anti-Submarine Warfare Officer and operations analyst before moving to the U.S. for my Ph.D. in Industrial Engineering. I've traveled extensively, and I’m now a proud U.S. citizen who enjoys philosophical conversations, animals, and diving into the latest in ML and neural networks. I’m also an active member of the Machine Learning Meetup groups, where we read and discuss research weekly.For me, optimization and data science go hand in hand — both aim to make the world more efficient, thoughtful, and intelligent.