Sven Balnojan

Hi! I'm Sven.

I'm a product manager & data developer in the realm of data. I like complex things, connecting business & technology and the B2B markets. I try to teach by writing about things I care about. I share my knowledge through blog posts, my data newsletter “Three Data Point Thursday” and some open source on GitHub. Read on, if you’d like to know more...

Writing & Talks

I write on Medium and sometimes hold a talk/workshop.

Newsletter #ThDPTh

I publish a data newsletter called "Three Data Point Thursday".

GitHub

I enjoy writing open-sourced things on GitHub. Mainly implementations & examples of more complicated technical things, workflows, technologies etc.

Technical/ Data Product Management

Currently I work as data product owner and am writing a lot about product management with a focus on machine learning and data products. Examples are “A checklist to identify a machine learning opportunity” or “The three product owner no’s” . Previously I was involved in an early startup in the Business Intelligence sector and worked 5 years part time in a B2B marketing agency.

Data/ML Development

In the data realm I care both about Machine Learning and broadly Analytics. My focus is on larger topics like architecture, organizing processes and bringing software development best practices to the data world.

Some details:

I’m an advocate for data decentralization, see my articles on the data mesh paradigm (more than one kind of data mesh, building a data mesh).
I think a lot about data processes, like how to organize ML teams and how to organize data teams.
Software dev. techniques I recommend to data teams as test-driven-development, trunk-based-development and many more.
On the machine learning engineering side I have some interest in geometric deep learning, and particulary for graphs, e.g. in graphCNNs , or in graph embeddings. And new technologies like Metaflow, and FKlearn.

DevOps Enthusiasm

I briefly worked as a DevOps guy in a data context, and what I learned impressed me. I still hold Infrastructure as Code Workshops and advocate for CI/CD or CD4ML in the machine learning world, ./go scripts, testing, testing, testing and many more techniques that are so essential to productivity in modern engineering teams. (And yes, I think every data team should work just like any other engineering/team)