CIVA is a Swedish physical AI startup building intelligence for machines operating in the real world.
We’re looking for a Research Engineer to advance how AI learns, represents and reasons about physical environments. You’ll work across model development, training and evaluation, turning research ideas into working systems.
Responsibilities
• Build learning pipelines for sequential machine, sensor and operational data.
• Develop and train models for representation learning and action-conditioned prediction.
• Design experiments, establish baselines and evaluate performance across operating conditions.
• Identify data gaps, model limitations and failure cases.
• Work with operators and industrial partners to connect model predictions with practical operating decisions.
Qualifications
• Strong Python and PyTorch skills.
• Experience training and evaluating deep learning models.
• Experience in one or more of: temporal modelling, multimodal learning, self-supervised learning, learned dynamics or robot learning.
• Ability to take a research question from hypothesis to implementation and evaluation.
• Evidence of work you’ve built: research, code, trained models or deployed systems.
You’ll help shape CIVA’s modelling approach, data requirements and evaluation methods from an early stage.
Interested? Send us your background and examples of your work.
