Visma Smartscan is a market leading document data extraction engine - processing more than 4 million documents every month - delivering automation to products like Visma eAccounting (SE, NO, FI, NL), Visma e-conomic (DK), Tripletex (NO), Visma Expense (NO, SE, DK, FI), Dinero (DK), Proactive (NL) and Yuki(NL)

Visma AutoSuggest simplifies the handling of more than 30 mio transactions per month within Dinero, e-conomic, Fivaldi (FI), Netvisor (FI), Proactive, Nova(FI) and more

The ML Assets team that builds these APIs is growing - and we’re looking for skilled software engineers with relevant experience to join the team.
The Job
As a software engineer in the ML Asset teams you will

Work on scaling and maintaining the inference APIs our customers rely on as well as the data platform they are based on.
Extend an improve the developer tools we provide to our integrators - APIs, auxiliary SDKs and the developer web portal with access to QA tools and integration setup
Help the data scientists on the team scale and automate our ML training pipelines.
Explore new applications and workflows exploiting our ML stack with our customers

We practice devops and automate everything that it’s possible to automate.

The Team

Curiosity, collaboration and a shared desire to change and improve what we know and how we work - that is what drives the ML Assets team.

We have an exceptionally high degree of freedom to innovate and experiment. We firmly believe the best products are created by highly engaged colleagues with freedom to operate.
We work with a research mindset and take the time to study the state of the art in our field. We keep our platform current and have a strong bias against technical debt.

The team is mostly based in Copenhagen - we’re a team of 1 product manager, 3 engineers and 3 data scientists. We’re now building out a presence in Košice.

The Stack

We’re big fans of Google Cloud Platform. Our APIs run on a Kubernetes cluster and are authored in Go. Our data lives in Google BigQuery. Our ML platform is based on Tensorflow and TFX - and the Python data science stack. We use Google’s Dataflow and the Vertex AI platform as much as we can.