Job Title: AI Engineer
Location: Stockholm
Employment Type: Full-time
Location Type: Hybrid
Department: Engineering

Compensation:

• Stockholm (Level 5): $126.5K – $149.5K • 0.23% – 0.69%
• Stockholm (Level 4): $104K – $123.5K • 0.19% – 0.57%
• Stockholm (Level 3): $88K – $104K • 0.16% – 0.48%
• Stockholm (Level 2): $77K – $91K • 0.14% – 0.42%
• Stockholm (Level 1): $71.5K – $84.5K • 0.13% – 0.39%

We believe transparency builds trust, and trust is what makes great teams work. That’s why all salaries at Filed are openly shared across the team and with every candidate. We don’t negotiate compensation; instead, every candidate chooses from three preset packages—ranging from high salary/low equity to low salary/high equity. We review levels every 4 months and adjust salary and equity based on performance and growth.

Responsibilities

• Designing and implementing LLM-powered workflows across tax, finance, and compliance
• Building and scaling Graphiti-like knowledge systems for memory and retrieval
• Owning prompt pipelines, vector stores, fine-tuned models, and evaluations
• Becoming an Oracle with a model of the customer and the domain in your head

Required Hard Skills

• Deep understanding of how LLMs work under the hood—beyond the hype
• Fluency with APIs like OpenAI, Gemini, Claude; experience with structured evaluations
• Comfortable working across Python and cloud infrastructure (GCP or Azure)
• Strong grasp of embeddings, RAG, structured retrieval, and production-grade inference
• Bonus if you’ve contributed to or built internal tooling for model debugging or evaluations

Required Soft Skills

• Ruthlessly pragmatic—knows when to ship and when to iterate
• Empathetic with users and critical of vague assumptions
• Deeply collaborative across product, engineering, and operations
• Obsessively curious—you want to understand everything about the domain you’re automating
• You own your outcomes—and love having your work in front of real users daily

Ideal Candidate

• Thinks in probabilities, systems, and second-order consequences
• Navigates ambiguity across LLMs, APIs, and semi-structured corpora
• Gets excited by the edge cases most people run away from
• No research-for-research’s-sake—ships fast and learns even faster
• Big bonus if you’ve shipped AI products into production, not just demos

Growth and Levels

We operate on six levels, from Level 1 to Level 6. Each reflects increasing ownership, impact, and scope with meaningful jumps in salary and equity. Promotions are based on output and ownership, not tenure.

Application Process

• Quick intro call
• Code and system walkthrough + async technical task
• Final round with our team
• Optional 2 to 3 day paid trial
• Then we decide together if this is the right fit

If you want to work with real users, real constraints, and build AI that goes far beyond chatbot demos—this is it.