The posting
About the role
We're now looking for an AI SW Engineer to focus on Agentic AI to build the agentic layer of the systems we ship. At Renessai we build AI-native software for clients: systems where agents do real work, not systems with a chat window bolted onto the side. We are looking for an AI Software Engineer whose home ground is the agentic layer, meaning tools, context, memory, orchestration, and the evaluations that tell you whether any of it actually works.
Two things define this role. First, agents are what we build: the core of most client systems we deliver is an agentic workflow. Second, agents are how we build: we develop with coding agents, skills, and orchestrators as the default way of working, and part of your job is making that way of working faster and more reliable for everyone around you.
Why Renessai
You'll be at the frontier of AI-native software, building real systems for real clients where agents are the core, not an afterthought. You'll own the work end-to-end and work in a team that develops with agents as the default way of working. Small company, fast decisions, serious impact.
You will ship early and learn from real use. That means rapid discovery, small real versions built to test a hypothesis, and a willingness to drop last week's idea when the feedback says so. It also means working close to the client and the people who use what you build, rather than at a comfortable distance from them.
What you will do
Build agentic systems as the core of client software Design and build agentic workflows: tool design, context engineering, memory, planning, and orchestration, including multi-agent setups when the problem genuinely calls for them. Decide where the agent ends and deterministic software begins. Not everything should be an agent, and knowing the difference is a large part of the job. Design the guardrails and human oversight a system needs to be trusted in production, and treat cost, scalability, and trustworthiness as design questions rather than things to fix later.
Build and run the evaluations Set up evals that test quality, catch regressions, and track performance against cost: accuracy, drift, correction rate, quality gating. Treat evals as part of the system, not a report produced at the end. If we cannot measure whether the agent got better, we do not know whether it did.
Build the way we build Set up the skills, harnesses, subagents, and context a project needs, and shape a development workflow where agents do a large share of the implementation and humans stay firmly responsible for the result. Keep improving that workflow as models and tooling change, which at the moment happens roughly monthly.
What you bring
You do not need every item below. You should be confident in several and willing to grow into the rest.
Agentic engineering Hands-on experience building agentic systems: tool design, context, memory, planning, orchestration. Your opinions on these come from things breaking in real projects, not only from reading about them. Practical experience building and running AI evaluations, and the judgement to weigh quality against cost. Daily fluency with coding agents (Claude Code, Codex, Cursor, or similar) that goes well past prompting: skills, subagents, parallel runs, and a working sense of what to delegate and what to verify yourself.
Software engineering Solid experience with at least one major cloud (Azure, GCP, or AWS), including setting up infrastructure, access, and governance. Enough backend and frontend experience to take a real system from architecture to interface without needing each piece handed to you as a finished spec. Practical experience with data pipelines: moving and synchronising data, including at larger scale.
Consulting skills and mindset Comfortable being close to the client and the people who will actually use what you build. Able to read a client's business context and users' needs well enough to make good calls without a full specification in hand. Willing to ship early, test with real users, and change direction when new information tells you to, rather than defending a plan. Tolerant of ambiguity and of things being unfinished. Straightforward and easy to work with, in a small, evolving company.
Practicalities
Location: You must be based in Finland and have a valid work permit in Finland. Language: Fluent English. Finnish is considered a plus. Office presence: This is a hybrid role. We expect you to, if necessary due to client work, have the possibility to come to the office or client site. Client site presence is subject to project and client needs. Other than that we are flexible and open to discussion of the cadence of physical presence at the office. Salary: We operate in set ranges per seniority level, these ranges will be shared with you early on in the hiring process.
How to apply
We're building a team with a genuinely wide range of backgrounds and ways of thinking. If you're on the fence about whether you fit, we'd rather you applied and let us make that call together.
If this sounds like you, send us your CV and we're happy to tell you more! We review applications continuously.
