Projects

Featured Projects

IoT ML Solutions & Logistics Optimization

Heineken (Freelance) | 2025 to Present

IoT ML Framework

Technologies: Python, IoT Sensors, Machine Learning, Optimization Algorithms

Delivering production ready ML projects in the IoT domain and logistics optimization space. Building effective frameworks and scalable code solutions.

Key Deliverables: - End to end ML pipelines for IoT data processing - Logistics optimization frameworks - Production deployment with monitoring and drift detection

Internal Knowledge Retrieval System

Rabobank | 2023 to 2025

RAG System

Technologies: Python, LangChain, RAG, ChromaDB/FAISS, OpenAI/HuggingFace

Developed RAG based system for information retrieval from knowledge bases supporting business decisions. Worked through multiple iterations of optimization and evaluation as independent contractor.

Delivered successful PoC, solution deployed to production.

Customer Retention Modeling

Rabobank | 2023 to 2025

Customer Retention Model

Technologies: Python, XGBoost, A/B Testing, MLOps

Developed predictive model to identify opportunities for proactive actions. Created, optimized, and tuned solution with comprehensive A/B testing framework as independent contractor.

Delivered successful PoC validated through pilot experiment.

Media Mix Modeling Solution

Heineken | 2022 to 2023

Media Mix Modeling

Technologies: Python, Bayesian Modeling, Causal Inference, Marketing Analytics

Managed interdisciplinary team of 15+ direct reports (6 data scientists, 2 front end engineers, 2 data engineers, 3 MLOps engineers, and others) and served as technical lead to deliver advanced Bayesian modeling solution for marketing optimization.

Impact: One of Heineken's most successful analytics projects, captured significant business value at tens of millions scale in the first year. Still actively used for strategic marketing decisions.

Financial KPI Forecasting & Time Series Best Practices

Heineken | 2019 to 2022

KPI Forecasting

Technologies: Python, Time Series, Statistical Modeling, Business Intelligence

Built from scratch with full end to end ownership: model architecture, development, MLOps, data engineering, project management, stakeholder management. Served as technical lead. Not just a forecasting system, but established time series best practices that became the foundation for the organization's approach.

Impact: Team later took ownership and evolved it. Used to this day for global financial planning.

Equipment Spare Parts Allocation

Heineken | 2019 to 2022

Spare Parts System

Technologies: Python, Optimization, Machine Learning, Data Engineering

Initiated and built with full end to end ownership: model architecture, development, MLOps, data engineering, project management, stakeholder management. Served as technical lead working with less senior colleagues.

Impact: Became a driving force for operational improvements across global operations with massive opportunity. Still live and used daily across the organization.

Full circle moment: Later returned as independent contractor for various improvements, proud and humbled to continue this work.

Learning Through Variety

Heineken | 2019 to 2022

In a fast paced, startup minded analytics environment, I worked on diverse projects to figure out what works in production ML. Some became tools still used today, others taught valuable lessons about real world constraints:

  • Equipment Failure Detection (IoT) - Explored real time anomaly detection for manufacturing equipment, learned the practical challenges of production ML systems
  • Data Quality Automation - Built tools for automated validation that are still in use
  • Bottle Design Recommender - Explored ML applications in product design
  • Warehouse Bin Allocation - Optimization experiments for space utilization
  • Delivery Route Optimization - Logistics ML and route planning experiments
  • Churn Prediction Modeling - Customer retention analytics
  • Workshops & Trainings - Shared knowledge and best practices across all organizational levels