Eindhoven, North Brabant, Netherlands
Building a product to help e-commerce businesses manage their inventory.
Building a product that optimizes and automates inventory for e-commerce businesses.
Developed an AI-powered app that writes structured documentation and database entries based on voice notes and sends them to the CRM.
The logistics department of bol.com is rapidly growing. The addition of new warehouses leads to challenges in daily capacity steering. I help by developing and deploying algorithms and machine learning models to optimize logistical performance, especially during peak (holiday) season. In my first nine months, I - developed an algorithm to balance order intake over the warehouses and transporters based on real-time capacity data - evaluated it using simulations and discussions with stakeholders about KPIs and what-if scenarios - deployed it to Google Cloud as the first Python-based microservice in the bol.com landscape (FastAPI, Pydantic) - added monitoring, logging and alerting to the application (Grafana, Kibana, Prometheus) - developed a dashboard to track the algorithms behavior and performance in real-time (Streamlit) After that I worked on a short-term workload forecast based on which interventions are made by an operational team. Some of my contributions are: - improving the code quality, structure, and accuracy of a Python-based simulation tool that simulates how customer orders are sourced from different warehouses and distributed in bol's distribution network - designing, developing, and deploying a machine learning application that predicts several sales composition characteristics for 'today' based on actual orders - designing a mathematical optimization model that outputs a set of customer orders to represent the predicted sales composition for use in the simulation - improving the team's capabilities for validation and evaluation of the forecasting tooling
In this freelance position, I work in the data science team at the Detecting Financial Crime (DFC) department on a large-scale Know Your Customer (KYC) project. Banks are under pressure of the Dutch National Bank to remediate (commercial) clients. At ABN AMRO, this means new processes are invented and resources are deployed to collect, organize and analyze data of 600k commercial clients. This is how I help: The first 6 months - I developed and deployed data pipelines in Python and Luigi to bring raw data from various sources together into usable formats and datasets. - I developed and deployed dashboards and other Business Intelligence functionality to provide insight into these new, ad-hoc processes. - I professionalized and streamlined the way of working of the data science team, e.g., by introducing dev-ops and automating recurring tasks. The remaining time, I defined and developed AI use cases focused around intelligent automation of the client review process. Specifically, - Object Detection and Classification in image-based pdf documents using both traditional Computer Vision techniques (OpenCV) and Deep Learning (Tensorflow, Keras, YOLOv3). - Image noise reduction and preprocessing for Optical Character Recognition to extract text from images and image-based pdfs (OpenCV) - Natural Language Processing to extract and classify relevant information from the extracted text.
JADS has various Professional Education programs in which industry professionals come to the university to learn about data science and how they can apply it in their business. As a "practitioner" for these programs, I did the following: - teach machine learning to groups of professionals - prepare course material and case studies - mentor and assess participants on their graduation projects
The Engineering Doctorate (EngD, formerly PDEng / Professional Doctorate in Engineering) program Data Science is a two-year post-master's program. It qualifies students with an MSc degree in mathematics, statistic and computer science to become top-level professionals. Such professionals help industry and business with their decision-making processes based on real, actionable data or become innovative entrepreneurs in the data science ecosystem.