{$ msg.text $}

Machine Learning Master Thesis

  • Applications are considered on a rolling basis
  • Applications are considered on a rolling basis

Accelerate the world’s transition to a clean, affordable and reliable energy system. This is our quest and vision at Greenlytics. We are building the software that will be the brain allowing for tomorrow's renewable, distributed and smart energy system.

Uncertainty in the power production is one of the key factors holding back accelerated integration of renewable energy into the power grid globally. We make difficult operation and dispatch decisions easy and reduce uncertainty for our customers. This is done by unlocking the latent value in the weather, market and production/consumption data, transforming it into actionable insights. That’s exactly what Greenlytics is all about.

Our technology is based on machine learning and artificial intelligence algorithms that forecast, optimize and control distributed energy resources to increase their value in the power system and market.

What have we accomplished?

Since the incorporation of Greenlytics in 2018, we have developed an energy forecasting algorithm based on artificial intelligence that gives state-of-the-art results when tested against competition in vendor trials. The forecasting service is currently running in live mode for our customers.

What is the challenge?

Because of the intermittent nature of consumption and renewable energy in the power system, it is important to forecast and plan it on a day-ahead basis. As opposed to deterministic forecasts, probabilistic forecasts provides uncertainty information that is essential for decision making in the power system and market. The goal of this master thesis is to evaluate and develop different probabilistic forecasting models using machine learning (specifically deep learning and decision trees) with a focus on wind and solar power production forecasts.

The necessary data to conduct the master thesis will be gathered from open machine learning competitions and from Greenlytics’ unique datahub containing easily accessible weather and production data. Working with open competition datasets will allow you to benchmark your results with researchers and data scientists from all over the world. Using the Greenlytics datahub will allow you to test your models on real world datasets while not getting stuck with time consuming big data engineering task.

Who are you?

We see that the student has an background with programming in data science languages such as Python, Julia or R. Having experience with neural network and decision trees frameworks such as Tensorflow/PyTorch/Keras and XGBoost/LightGBM/CatBoost is a plus. Furthermore, the student should have a good understanding of project relevant subjects such machine learning and statistics.

Benefits

We offer the possibility for a master thesis student to translate their theoretical skills to solving a real world problem that has a large impact potential on the climate change. The student will work in close cooperation with data scientists and domain experts at Greenlytics and our partner research institute (RISE SICS) as well as getting experience from working in a startup environment.

About the company

We at Greenlytics are on a quest to develop the software that will be the brain allowing for tomorrow's renewable and distributed energy system.

By unlocking the latent value in the data and transform it into actionable insights, we make hard decisions easy for our customers. This is achieved using algorithms that builds on machine learning that allow to forecast, optimise and control distributed energy resources to increase their value in the power system and market.

Sebastian Haglund El Gaidi | Hiring Manager

I'm interested