Naar inhoud
About us

Graduation Assignment - Classify Products With Machine Learning

Graduation Assignment in the area of: Computer Science - Data Science

Classify Products with Machine Learning


Pincvision advises and supports multinational companies with all their compliance processes arising from cross-border goods transport and will even handle the administrative aspects of these activities. Thanks to its years of experience, expert knowledge, state of the art software applications and a worldwide network, Pincvision is the ideal partner for global companies. Through our services several Fortune 500 companies already experience that compliance is no barrier to global trade.

We bring together the best specialists in laws and regulations relating to VAT, Intrastat, Environmental compliance, Customs and Export Documents. Our business operations are supported by innovative software solutions involving the smart application of decision rules and cross checks. These solutions guarantee consistent quality, ensure that large volumes of data are processed rapidly and help optimize our customer’s processes as a result.


When preparing and submitting declarations for VAT, Intrastat and Environmental, we process large numbers of both financial and logistic transactions. This are data sets and reports retrieved from our customer’s SAP, Oracle or other ERP systems, but also from external data sources. On a monthly base we process between 1.5 and 2 million transactions for our customers. The total value of all these transactions on a yearly base is over a 100 billion Euro.

Currently, we focus on an efficient and flawless process for processing raw data into regular declarations which we submit with international authorities. For the Intrastat (inside EU), Environmental en Customs declarations products needs to be classified with a Harmonized Tariff Code (HTS) code to categorize products. This is already delivered in the dataset of the customer or we determine this during as part of the classification service. The HTS code is defined based on all kind of product characteristics. An example of an incorrect classification is classifying a real vessel under a HTS code that reflects a toy vessel. This mistake makes a huge difference in terms of import duties when the vessel is being import cleared at customs.

Pincvision wants to investigate how classification of large number of products can be done in an automated way with the use of some form of artificial intelligence. The main question is whether machine learning techniques can help us to classify/verify the HTS code?

This can help up with the following points:

  • Improve quality of classification process
  • Automated validation of received HTS code in datasets
  • Increase customs process and prevent delays and audits

Technical challenges:

  • Research on available solutions and algorithms to find the best fit to validate or classify products
  • How to check the reliability of the classified products?
  • Integrate results with existing applications on Microsoft Azure


To be able to address these and other questions your graduation assignment will cover a number of research questions to define an approach and the realization of a proof of concept to show how this is working in our environment. The scope will be further detailed in mutual agreement. Preferably the following topics will be covered to a greater of lesser extent:

  1. What type of approach and tools, techniques to conduct machine learning are currently used and which will fit the best for Pincvision
  2. What type of tooling is most widely used and which will fit best for Pincvision
  3. Do we need to add additional product characteristics to improve the reliability of the machine learning results
  4. Develop a prototype of the solution, preferably that integrates with the existing tooling within Pincvision that is based on a Microsoft Azure environment


  • HBO/WO education in the area of Data science, Business Intelligence (BI), Statistics, Engineering, Computer Science or Mathematics
  • Able to translate ambiguous business problems into a conceptual mathematical architecture
  • Interested in data science, data analytics and latest IT developments in this area
  • Hands-on experience with a wide variety of data mining, predictive analytics, machine learning
  • Skills: problem solving, critical thinking, self-reliant, and clear communication
  • Experienced with SQL Server databases, Machine learning, Python, R and/or other relevant tooling


Are you interested in data science, data analytics and latest IT developments? Quickly respond to this unique opportunity! Apply directly by filling in the below application form. Please upload your CV and letter of motivation.

For more information about this vacancy please contact Edwin Kampshoff, CIO,

Apply for this job!

Fields with a * are required


Pincvision uses cookies to ensure an optimal userexperience.

We would like to ask your permission to place cookies.

More information.

Accept Decline