Andreas Hellander
Co-Founder / CEO
Research roots.
Operational focus.
Scaleout builds software infrastructure for AI across devices and sites. We bring distributed machine learning research together with the engineering needed to put it to work.
Why we buildBring learning closer
Bring learning closer
to where AI operates.
Useful AI has to work with the data, hardware, and connectivity available in the deployment. The observations that could improve a model often cannot be collected in one central place.
Our work connects model operation, local improvement, and learning across sites. Scaleout Edge provides the shared foundation; application software and engineering connect it to the task.
Founded by researchers from Uppsala University, Scaleout combines experience in distributed systems, scientific computing, and machine learning.
People
The team behind the work.
Researchers, engineers, and people building the organisation around them. Our work spans machine learning, software infrastructure, and integration in the field.
Salman Toor
Co-Founder / CTO
Daniel Zakrisson
Co-Founder / COO
Jens Frid
Co-Founder / CMO
Ebba Kraemer
Co-Founder / CCO
Fredrik Wrede
Head of Engineering
Mattias Åkesson
Senior FL Engineer
Stefan Hellander
CISO
Niklas Holmström
Software Developer
Viktor Valadi
ML Engineer
Jonas Frankemölle
ML Engineer
Katja Hellgren
ML Engineer
Carl Andersson
ML Engineer
David Hovstadius
ML Engineer
Sigvard Dackevall
ML Engineer
Muhammad Usama
ML Researcher
Rasmus Maråk
Research Engineer
Lucas Beerens
MLOps Engineer
Robin Kurtz
Senior ML Researcher
Srihaarika Vijjappu
ML Engineer
Magnus Blåudd
Principal Software Developer
Noah Wassberg
Research Engineer
Ragnar Hägg
Data Engineer
Svea Ottestig
Operations
Careers & research
Work on what comes next.
Interested in distributed machine learning, edge software, or the engineering around real deployments? We’d like to hear from you.
Join the conversation.
Tell us about your experience and the kind of work you want to do. For general career enquiries, get in touch with the team.
Contact ScaleoutMaster’s thesis projects
Explore the project descriptions for research questions and prerequisites. See each project description PDF for further details and application instructions.
Communication-efficient federated learning
Investigate dynamic quantisation-aware training and the balance between model accuracy and communication costs.
Read project description (PDF)Context-aware mixture of experts
Explore routing, quantisation, and pruning for efficient vision models on constrained devices.
Read project description (PDF)Heterogeneous federated learning
Study adaptive sub-networks for clients with different compute and memory resources.
Read project description (PDF)Domain adaptation with mixture of experts
Investigate adapting to new domains while retaining knowledge from earlier tasks.
Read project description (PDF)Security in federated object detection
Study active gradient inversion and the trade-offs between privacy protections and detection performance.
Read project description (PDF)Efficient vision through distillation and adaptive deferral
Investigate compressing both vision models and input resolution using knowledge distillation. Explore confidence-based deferral to a larger model for uncertain predictions, and assess the trade-offs in accuracy, latency, and energy use.
Project description PDF