Senior Software Engineer (AI-native product engineer)

Overview Swarmia was founded to help engineering organizations ship exceptional software at a sustainable pace. As Swarmia engineers, we’re not only building a platform for our customers but also ourselves. We raised €10M in Series A funding in June 2025 and are growing fast. To help scale and expand the product, we’re looking for several Senior Software Engineers who are both generalists and specialists: you know a little about everything and a lot about something. About Swarmia Swarmia is a SaaS product that gives engineering organizations the visibility they need to ship faster without compromising culture or quality. Our customers include some of the world’s best software companies, such as Miro, Superhuman, and Bolt. We’re a team of experienced engineers who care deeply about our craft. We practice what we preach: we use Swarmia to build Swarmia, and we’ve figured out how to ship fast while keeping technical debt low and our work-life balance healthy. About the role and how we work You’ll work as an AI-native product engineer across the full stack - from data models and APIs to UI - owning open-ended customer problems end-to-end. That often means starting with a working prototype rather than a finished spec, setting up the conditions for agents to do high-quality work, and making the judgment calls that agents still can’t. We practice what we preach. Our founder and CEO Otto Hilska described the shift this way in The New Way: Building software in an agentic organization: “The job of an engineer used to be mostly about delivering a product. The new job is to deliver a system that delivers a product.” Engineers use AI coding tools and agents as a default part of the workflow, maintain the repository context and skills that make agents effective, and review both human and agent-generated code with an eye on quality, blast radius, and long-term maintainability. Every Friday we demo what we’ve shipped. The culture is continuous shipping at a sustainable pace, high ownership, and low bureaucracy. No made-up deadlines, solo-projects, or needless bureaucracy. We value initiative, ownership, transparency, feedback, and collaboration. When you see something that slows the system (flaky tests, missing agent context, unclear ownership, slow feedback loops), you fix it or start the conversation. Process exists to protect quality and flow, not to slow people down. We have enough process not to worry about the basics, but total freedom to influence what matters. Examples of things you’ll do Take open-ended product problems (not fully specified tickets) and turn them into working prototypes with agents in hours or a couple of days, then use the prototype to decide what to cut, rewrite, de-risk, or ship. Collaborate live with product managers, designers, and other engineers in short joint sessions (Campfire-style) where the group shapes and iterates on real code rather than sequential handoffs. Design and implement the system that makes agentic work reliable: repository instructions (AGENTS.md and skills), fast and trustworthy CI/tests, clear boundaries for what agents can own, and automatic handling of routine work (dependency updates, certain exceptions, boilerplate). Build features end-to-end - database schema and GraphQL API changes, React components guided by the design system, integrations with GitHub, GitLab, Jira, Linear, and AI tool data - then split the work into reviewable pull requests. Review at least as much code as you write. This includes interrogating agent-generated diffs (focusing on tests, trade-offs, security, and architecture rather than line-by-line reading), codifying repeated feedback into agent skills, and keeping review from becoming a bottleneck. Help the team delegate more complex work to agents over time while maintaining quality. Debug slow queries, write the occasional complex SQL, jump into customer support chats for technical issues, and fix things that bug you before they become bigger problems. Contribute to the product’s own AI and agent-related capabilities (and use Swarmia on Swarmia) so the platform stays ahead of how engineering organizations actually work. Write the occasional blog post about something you’ve built or learned, and help teammates get unblocked. What good looks like Proven experience shipping product with modern AI coding tools and agents (Cursor, Claude Code, Copilot, Codex, or similar). Comfortable prompting end-to-end prototypes, iterating on agent output, and knowing when to take over. Ability to create and maintain the conditions agents need: clear repository context, skills/AGENTS.md files, reliable automated tests and CI, and sensible boundaries so agents can own meaningful work. Strong product sense: you can take a customer problem and constraints, prototype solutions, and make (or recommend) the call on what to build. You’re comfortable owning outcomes, not just implementing specs. Judgment under higher throughput: you review agent-generated code effecti