Principal AI Engineer

Location: Gothenburg, Sweden

Employment type: Permanent

Compensation: ~1,041,500 SEK base salary + bonus (total package ~1,149,500 SEK)

The opportunity

We’re supporting a global, research-driven organisation that is scaling how advanced AI systems are built, deployed, and governed across critical production environments.

This role sits at the centre of that effort.

This is not a pure management role and not a research-only position. It’s a senior, hands-on AI engineering role for someone who wants to stay close to complex technology while also shaping standards, architecture, and how AI is applied at scale.

You’ll work on real-world, production-grade AI systems, including modern transformer-based models, with real expectations around robustness, scalability, security, and long-term maintainability.

What you’ll be doing

• Acting as a senior technical authority for applied AI and machine learning systems in production
• Designing, reviewing, and improving transformer-based and ML systems used in real-world applications
• Remaining hands-on with models, training pipelines, evaluation, and failure analysis when required
• Providing technical leadership and mentorship to senior engineers through deep involvement rather than delegation
• Influencing architectural decisions around model selection, fine-tuning strategies, deployment patterns, and infrastructure
• Working closely with platform, engineering, and product teams to translate requirements into reliable AI solutions
• Helping define best practices for training, fine-tuning, monitoring, and operating AI systems at scale

What we’re looking for

We’re looking for someone operating at Principal / Staff / Lead Engineer level who is comfortable going deep technically and taking ownership of complex AI systems.

You should bring:

• Strong, hands-on experience with applied machine learning and modern AI systems
• Deep understanding of transformer architectures and attention mechanisms, including:
• How transformers process sequences
• What attention is doing under the hood (queries, keys, values, multi-head attention)
• Ability to reason about how individual tokens move through an attention layer
• Solid understanding of embeddings and positional information, including how transformers handle word order despite attention being order-agnostic
• Practical experience fine-tuning pretrained models, including familiarity with modern parameter-efficient approaches (and why teams choose them in production)
• Real experience training and operating large models, with an understanding of:
• GPU memory usage
• Where memory is consumed during training (parameters, gradients, optimizer state, activations)
• Strong grasp of normalisation techniques, including:
• Differences between batch normalisation and layer normalisation
• Why certain approaches are standard in transformer architectures
• Experience running ML systems in production environments, not just experimentation or research
• Familiarity with cloud and/or hybrid infrastructure, containers, and ML pipelines
• The ability to lead technically, review others’ work critically, and step in when systems fail

Previous people leadership or mentoring experience is valuable, but technical credibility and hands-on depth are essential.

Why this role

• Work on meaningful, real-world AI systems rather than proofs of concept
• Strong technical culture with high expectations and high trust
• Clear long-term investment in AI as a core capability