Achieving Productive Data Systems

From the Korfball court to the mountain sides, Jelmer has learned to teach by intensive action and practice. An eclectic array of training and teaching styles across a collection of hobbies and sports inspired him to help others as many helped him. At Aurai, Jelmer’s task is to equip others with the basic tools to build digital information into productive systems with useful interfaces.

“Machine learning engineering is all about bringing data science projects into production. It includes data collection, data validation, optimization of models, automatic training pipelines, serving the model and monitoring its performance. Data engineering is about supporting other roles in their data projects. Data engineers create data ingestion pipelines, data processing workflows, set up databases and help with end results such as visualizations and dashboards, but also implementations via APIs.”

Teamwork and Diverse Collaboration

Jelmer enjoys data engineering and training over pure programming because it requires teamwork and more diverse collaboration. “Good programmers should be able to explain their ideas and work closely with other people, such as colleagues and stakeholders. It requires interaction, so does training. It’s great to always learn more about active education, didactics and how to best help people learn.”

Improving Data Systems

Even with prolific experience in the data software industry as a business entrepreneur, freelancer and academic, he wanted to dive deeper into practical data science application and joined Aurai’s Data Specialist traineeship. Every traineeship bridges companies to trainees to model real systems and services. For Jelmer, it was a launching pad to help companies improve their data infrastructure and possibilities: “Over the different jobs and clients with Aurai, I have worked with many software stacks, almost all of them Python based. One was a C# based stack. My main focuses were data modeling and database setup, natural language processing and time series prediction.”

20% Theory, 80% Practice

Thanks to Jelmer, Aurai’s educational department is training data scientists and specialists to use and communicate the latest and most effective data engineering and machine learning tools. “Our traineeships last two months. Our machine learning engineering traineeship is around 4 people. Their goal is to be able to transform a data science prototype project into a fully functioning project in production. Our data engineering traineeship is around 6-8 people. Their goal is to know the basics of many types of tools (to meet any client need) and to build the mental framework to quickly gain skills in tools that are similar to what they’ve seen during the traineeship.”

“We learn a lot by doing tasks ourselves. This is consistent in the guided projects at the end of the traineeship where we simulate real projects by interacting with a client and creating a trial project for them. We structure the training on roughly 20% theory (presentations, reading) and 80% practical split (discussions, tutorials, projects).”

Providing More Choices Expands Possibilities

Each traineeship not only has new students but new clients with novel data challenges that demand new models. By teaching data engineering, Jelmer can also help trainees use the theoretical background on how our systems function to improve analytic and expansive thinking and choices in their business and personal lives. He hopes that training new data experts into engineering will expand the overall tech industry to help more workers and companies.

“Many of the trainees are junior data professionals that know some tools well, but they aren’t familiar with the background architecture considerations and choices and how to bring it all into a final project. I think by adding this knowledge we enable trainees to join in on the conversation of why and how earlier, and thereby provide data solutions representing well-thought-out and sustainable choices.”

Interested to see Jelmer in action and join Aurai’s traineeships? Find out more here!