Who we are
Enterprise work still moves by hand: copy-pasting between spreadsheets, endless email threads, and clunky legacy UIs. We started Duvo to end that for good.
We've already earned the trust of a range of customers, and our agents are helping them automate business-critical processes. We are growing fast, but to win from here we need exceptional people; that's where you come in.
What we are building
We're building the AI operations platform for large enterprises, currently focused on retail and consumer packaged goods customers. In Duvo, customers build AI agents that execute work wherever it needs to get done—SAP, spreadsheets, supplier portals, email, APIs, you name it. Duvo is heavy on browser and computer use.
In Duvo, business users specify the outcome; agents plan, act, request approvals on exceptions, and learn with every run. To help customers understand what to automate, we also help them map their processes by digesting interviews and internal documentation, which gets our foot in the door. We start with automating the parts of companies that we know best (category management, supply chain, finance ops) where we can show value fast, then expand to adjacent functions and sectors.
Velocity is our moat: ship fast, iterate faster, compound learning.
The role
You will own the AI infrastructure that makes our agents reliable, fast, and safe in production. You build the agent runtime, evaluation pipelines, context management systems, tool orchestration, and the observability tooling that lets agents execute end-to-end work for enterprise customers.
This is applied AI systems engineering, not ML research. You ship production systems that use LLMs, retrieval, and agent orchestration—and you're accountable for their reliability, cost, and quality in the real world.
Your unit of ownership: the AI platform layer — agent runtime, context management, tool execution, evaluation harnesses, and prompt engineering for system behaviors. You own sandbox behavior and agent runtime logic; SRE owns sandbox infrastructure and capacity.
We're a growing product team scaling into multiple initiatives, each with a lead, engineers, a design engineer, and an AI-focused engineer.
What we're looking for
These are non-negotiables—the things we'll specifically evaluate you on:
- Experience building production AI systems. Not research — making LLMs reliable at scale. You've dealt with prompt engineering, context management, tool use, or agent orchestration in production.
- Shipping and ownership. You've taken ambiguous AI problems to production with measurable outcomes. You own the full lifecycle — build, evaluate, deploy, monitor, iterate.
- System design for AI. You can design systems that handle the unique challenges of AI: non-deterministic outputs, context window limits, tool execution failures, and cost optimization. You think about reliability, cost, and latency as first-class concerns.
- Evaluation design. You can build evaluation frameworks that catch regressions, measure quality, and give the team confidence to ship AI features. You understand failure taxonomies and know how to create meaningful test sets.
- Debugging and diagnosis. You're hypothesis-driven when things break. You can trace failures across model behavior, data pipelines, and infrastructure to find root causes.
- Judgment as AI evolves. You'll make build-vs-integrate decisions on AI infrastructure with incomplete benchmarks, and course-correct fast as models and providers evolve.
You might also
- Have scalable, distributed-system instincts—you've designed and operated systems that handle high throughput and complex failure modes.
- Have a strong sense for security in AI systems—prompt injection, insecure output handling, supply chain risks.
- Have contributed to open-source AI tooling or infrastructure projects.
This is not for you if
- You want an ML research role — we don't train models.
- You primarily want to build user-facing product features and UI.
Location
We're looking for people within two hours of Prague (CET), so the team overlaps for most of the working day.
Our tech stack
- GCP (Cloud Run, GKE, GCS)
- Terraform, Docker
- Datadog
- TypeScript and Python services (you'll read and occasionally modify application code, but deep language expertise isn't required)
- Postgres, Redis
How we work
These are real tradeoffs we've made, not aspirations:
- Initiative-driven. We organize around customer problems, not org charts. Problems surface through product feedback, competitive analysis, and direct customer conversations — then we prioritize, build, and ship weekly.
- Customer-obsessed. We solve real problems, not hypothetical ones. Features that don't move customer metrics get cut.
- Iterative by default. We ship small, learn fast, and never get attached to yesterday's code. This means things break sometimes — we fix forward.
- AI-first leverage. We use AI to move faster and focus human time where it matters most. If a tool can do it, a person shouldn't.
- Direct feedback. We give each other actionable feedback immediately. This can feel uncomfortable — we think that's worth it.
- Autonomy with accountability. We trust people to make decisions and hold them to outcomes, not process.
What we offer
- Unlimited AI budget. We don't just allow AI tools — we strongly encourage them. Want to try a new tool? Buy it. Want to automate part of your workflow? Do it.
- Autonomy to do your best work. Want to meet someone to learn from? Set it up. Want a mentor? Go get one. Want to fly out to talk to an important customer? Just ask.
- A real AI product with real customers. You're not building demos or internal tools. Enterprise customers use what you ship, and their feedback drives what you build next.
- A sharp, motivated team that values ownership and candor.
How we hire
We respect your time and aim to move fast:
- Discovery call (online, 30 min). We'll talk about you, how you think and whether there's mutual fit.
- Technical interview (online, ~1 hour). Meet the team. We'll go deeper on your experience, system design, product thinking, and collaboration. No trick questions — we want to see how you think and build.
- On-site trial day (1 day). Ship something small to production with us and see how we work together. Compensated.
