Gaining insight and value from data requires a considerable commitment to and knowledge about data gathering, structuring and processing. Employing machine learning services demands skill in problem formulation, modelling and software engineering for deployment and integration. We deliver services on each step in these processes and aim to enable our customers digitalization strategy with our extensive knowledge of:

  • Data analytics and machine learning
  • Architecture and platforms
  • Customer training


Data analytics & machine learning

Machine learning applies powerful algorithms to potentially huge data sets in order to extract information, learn patterns and predict future behavior. This provides companies with high-quality information for making business decisions, task automation and predictions that often outperform human ability.

Machine learning problems

Here we present a few sub-fields of machine learning and example problems which they efficiently solve.

  • Regression predicts a real number: house price, expected product lifetime, tomorrow’s temperature, etc.
  • Binary classification answers yes/no questions: will the customer buy, is the e-mail important, will the customer leave, etc.
  • Anomaly detection finds atypical values: fraudulent transactions, production errors, sensor anomalies, etc
  • Recommendation systems make tailored recommendations: TV-shows and movies, tomorrow’s dinner, musical artists, etc.
  • Reinforcement learning trains autonomous agents: robotic control, elevator optimization, playing chess, etc.
  • Segmentation automatically groups data: customer segmentation, exploratory data analysis, etc.

We also solve optimization problems, in which there are typically no randomness. Examples problems include: shortest route, best utilization of machines and resources, pairing taxis with travelers, shift scheduling, shipping routing, etc.

Customer value

Data analytics, machine learning and artificial intelligence can help businesses with the following:

  • Improve internal processes and resource allocation.
  • Automate manual tasks.
  • Discover new potential customers.
  • Predict customer behavior.


Contact us to learn more about how we create value in current and previous projects.


Architecture and platforms

Once a model has been created it must interact with incoming data, adapt to new data and be available without interruption. In other to facilitate this, software architecture and data platforms are needed.

These systems are increasingly being moved to the cloud, and we offer expert help on popular cloud data platforms.

Why cloud platforms?


Cloud solutions are an abstraction of physical machines, freeing the user to focus on value creation instead of physical server layout, server administration, scaleability and security. Big data can be processed quickly, models can be served 24/7 without interruption, while the surrounding infrastructure is continuously guaranteed by the cloud provider.

We build solutions for all three main cloud platform providers, Microsoft Azure, Google Cloud, and Amazon web services. In addition, we have experience working with the Cognite Data Fusion platform and the integration between “on premise” and cloud systems.


Customer training

Full-day training courses in applied analytics, machine learning and data platforms. We frequently host customized courses catering to the interests of our customers.

Training course content

Courses by Sonat always introduce both the theoretical foundations, as well as practical exercises for attendants. A typical course covers the entire process from problem definition, data preparation, exploration, data modeling, model evaluation and deployment. This gives the participants insight and experience with fundamental methods to:

  • Retrieve information from data.
  • Predict future events and outcomes from existing data.
  • Evaluate how well a predictive model performs, and how to improve it.
  • Share insight and use models across the organization.

The difficulty will be adjusted based on the background of the participants. We include plenty of real-world examples, and if companies curious about whether or not machine learning can be applied to their problems will find that we are eager to discuss possible use-cases.

See our course product sheet for more information.


Sonat AI

Data analytics, machine learning and optimization.


Kjell Værøy Ljøstad

CEO Oslo
+47 952 67 129

Sonat Consulting Oslo AS
Karl Johansgt. 25
0159 Oslo


Jonny Klemetsen

CEO Bergen
+47 294 56 245

Sonat Consulting Bergen AS


C. Sundts gate 17-19

5008 Bergen


Sonat Consulting Bergen AS

Postboks 234 Sentrum

5804 Bergen