FEDERATED LEARNING ENGINEER

FEDERATED LEARNING ENGINEER
Our client SMARTERGY AB is a pioneer in AI predicted distributed Energy Storage Systems.
Mutual Benefits Engineering, MBE are growing and are looking for a senior Federated Learning (FL) Engineer to contribute to our client´s significant effort of delivering huge amount of electricity to cars, trucks, properties, and industry.
Our team take care of everything from migrating an extensive amount of data, with its ETL process, data, and reporting solutions to an AWS cloud environment to visualize on our APP and developing the hardware in distributed energy storage systems, ESS. The ESS contains high-end AI controlled energy storage system.
Roles and Responsibilities
· Contribute to the development of Federated Learning’s algorithms in our central server DEEP and at the devices.
· Evaluate and validate results in our testbed in Alafors, 20 min from Gothenburg.
· Contributing to and coordinating with SMEs regarding ESS equipment, ETL-process and customers feed back.
· Support stakeholders and partners in the transition to the new AI-tool, including support with queries, troubleshooting and training.
Mandatory Requirements
· Expertise in different Machine Learning algorithm programming, Python, gathering requirements, designing, and developing embedded systems.
· Experience working with and designing power electronic systems.
We expect you to have.
· A degree in relevant field
· More than 5 years of relevant working experience
· Expertise in Machine Learning technologies and preferably Federated Learning.
· Fluent English language skills
It would be great if you also possess.
· Experience with migrating data solutions into embedded systems
· Experience with programming Python languages.

In order to succeed in this role, you also need to be structured, analytical, and thorough in your work in order to handle development and evaluation of an extensive and new solution.

ABOUT Mutual Benefits Engineering, MBE
Advanced Technology and Management
MBE have solid experience of sustainable product development in Automotive and Clean Energy Tech Industry in combination with advanced AI for electrical grid (similar to stock market). After several awards, for technical expertise, from European Council, Swedish Energy Agency and Västra Götalandsregionen as well as for growth and leadership from DI Gazelle we feel confident in our progress and are welcoming additional experts to our journey.
Versatile & Inclusive team
We are happy to see applications from creative, curious and senior developers to join our expert team towards advanced technology clients within Automotive and Circular Energy. We are extra happy to see applications from FEMALE applicants.
We engage top engineers, experienced experts from PhD level as well as distinguished newly graduated engineers with personal values for Sustainable Engineering, AI and Power electronics, caring for a concrete better society, well-balanced life together with your best career development!
Apply, Develop and Create your Future
Join us on the travel to your best future! Send your application as soon as possible to jobb@mutualbenefits.se . Name your application ”Business Intelligence Expert”. We continuously review applications.


Federated Learning in our perspective
Federated Learning is a machine learning approach that enables training models across decentralized devices in our ESS’s while keeping data localized. Instead of collecting all the data in a central server, federated learning allows the model to be trained on the device where the data is generated. This approach has several advantages, including privacy preservation, reduced communication costs, and the ability to leverage data from multiple sources without centralizing it.

Here's a breakdown of the key components and the process involved in federated learning:
1. Initialization:
A global model is initialized on the central server, DEEP (Digitized Energy Efficiency Portal). This model serves as the starting point for training.
2. Distribution of Model:
The global model is sent to individual devices on each ESS where local data is stored. Each device performs model training on its local data.
3. Local Model Training:
On each device, the local model is trained using the locally available data. This step can involve multiple iterations or epochs to improve the model's performance.
4. Model Updates:
After local training, only the model updates (changes to the model parameters) are sent back to the central server, rather than the raw data.
5. Aggregation:
The central server collects the model updates from all devices and aggregates them to update the global model. This step is done using federated averaging techniques.
6. Iterative Process:
Steps 2-5 are repeated iteratively. The global model is sent back to the devices, local training occurs, updates are sent back, and the global model is updated.
The key advantages of federated learning include:
· Privacy Preservation: Raw data remains on the local devices, and only model updates are shared. This helps address privacy concerns associated with centralized approaches.
· Reduced Communication Overhead: Only model updates, which are typically much smaller than the raw data, are transmitted between the central server and devices. This reduces the need for large-scale data transfer.
· Decentralized Training: Federated learning allows for training models on data distributed across multiple devices, making it suitable for scenarios where data cannot or should not be centralized.
Öppen för alla
Vi fokuserar på din kompetens, inte dina övriga förutsättningar. Vi är öppna för att anpassa rollen eller arbetsplatsen efter dina behov.