Staff Applied AI Engineer

Job Description

Do you want to own the technical direction of an agentic AI platform serving product teams across Europe?

Are you interested in joining of Europe's fastest growing companies? Do you want to impact the wellbeing of millions of users?

If yes, join us to make an impact in a fast-paced and dynamic environment where you’ll play a key role in driving financial accountability, supporting strategic decisions, and enabling sustainable growth.

Epassi’s purpose is to boost everyday well-being. We are a leading provider of employee benefit solutions across Europe. We were established in Finland in 2007, and in 2008 we were the first company to launch a mobile-payable employee benefit payment solution in Europe. Since then, we have consistently grown, diversifying our products and introducing our services into new markets.

Epassi has been awarded by the Financial Times as one of the fastest-growing companies in Europe on multiple occasions.

Your Mission at Epassi:

You will set the technical direction for Epassi's shared AI platform: the agent runtime, LLM gateway, and tool-calling infrastructure that every product team builds on. Employees use our products to discover and use their benefits in natural language, and employers rely on them to understand which benefits deliver value. You decide how we architect for that across markets: which patterns become platform primitives, where the boundaries sit between the shared layer and product teams, and how we keep quality measurable as adoption grows. This is a hands-on Staff role: you set direction through working code, and you build the hardest parts of the platform yourself rather than delegating them.

Your responsibilities will be:

• Own the architecture of the agent platform on Amazon Bedrock and build its core components yourself: runtime design, tool registry, orchestration patterns, model routing, and the LLM gateway other teams call.
• Define the platform contract: which capabilities are shared primitives, which stay product-specific, and how teams integrate through APIs, SDKs, and reference implementations.
• Establish the eval and quality strategy for the whole platform: eval infrastructure, regression gates in CI, and quality metrics that leadership can read.
• Set the security and compliance posture for agents in a benefits and payments domain: prompt injection defense in depth, PII boundaries, data residency, and auditability.
• Make the platform economical: model selection, caching, and routing strategies that keep cost and latency inside explicit budgets as usage scales.
• Ship production code continuously: unblock the hardest problems in the codebase yourself, and write the ADRs that settle contested technical decisions.
• Mentor senior engineers across teams and raise the organization's bar for building with LLMs.

What You Should Bring:

• Experience: 8+ years of backend engineering in Java or Python, including systems that outlived your involvement because you designed them well
• Agentic systems in production: You have designed and operated LLM-powered systems at scale: agent loops, tool orchestration, RAG pipelines, or multi-model routing, and you have opinions grounded in incidents, cost curves, and eval data
• AWS depth: Strong AWS architecture experience; hands-on Bedrock experience is a strong plus, and deep production experience with Anthropic, OpenAI, or Opensource models APIs.
• Platform leadership: You have built infrastructure or internal platforms that multiple teams adopted, and you know that adoption is won through developer experience and reliability.
• Hands-on at Staff level: You have stayed a builder as you grew in seniority: your influence comes from working code, reference implementations, and production contributions, and you expect to keep coding most of your week in this role.
• Evals as engineering: You treat LLM quality as an engineering discipline: eval datasets, automated regression, and metrics that catch degradation before users do.
• Security judgment: You understand the threat model of agents with tool access to sensitive data: prompt injection, data exfiltration, and least-privilege tool design.
• Technical communication: You write ADRs and design documents that settle debates, and you can explain model behavior and trade-offs to executives without hand-waving.
• AI-native development: You work AI-natively yourself and you set the standard for how the organization uses tools like Claude Code, Cursor, or Copilot responsibly and productively.

Language Profile:

• Fluent in English (working language)

Why You Should Join Us:

Epassi is one of Europe's fastest-growing employee benefits companies, serving millions of users across the Nordics, DACH, and Southern Europe. You will join at the moment we build our AI platform from the ground up: the agents you ship will power benefit discovery for employees and analytics for employers across every Epassi product. You get real ownership, a modern AWS-based stack, an AI-native engineering culture, and colleagues across Europe who care about shipping.