Software Product Engineer (Agentic & AI Systems)
Remote-first (EU timezone) · Full-time
At FifthRow, we are building the future of AI consulting and analysis. Our platform delivers autonomous agents and workflow automation that turn complex strategy, research, and decision-making into software. We believe that workplace AI should do more than answer prompts - it should orchestrate, execute, and deliver measurable outcomes across entire business processes.
We’re looking for a product-minded engineer who can design, build, and iterate on agentic systems end-to-end: from problem framing, to architecture, to shipping real features used by customers - and sometimes by our own team.
What you’ll actually do
Build AI-native product features
Design and evolve LLM-powered workflows and agentic systems that solve real customer problems.
Build multi-step agents with state, memory, tool use, routing, and evaluation.
Turn fuzzy requirements into working systems through fast iteration and feedback.
Own features end-to-end: design, implementation, iteration, and reliability.
Ship like a product engineer
Collaborate closely on product direction and tradeoffs, not just implementation.
Think in outcomes, user value, and constraints-not just architecture diagrams.
Consider cost, latency, observability, and failure modes from day one.
Use AI to accelerate development itself: prototyping, refactoring, testing, evaluation.
Examples of problems you might work on
Designing an agentic workflow that turns unstructured inputs into structured, reliable outputs.
Building systems that adapt behavior based on context, usage signals, or prior interactions.
Improving an existing AI feature that works “most of the time” into one that works predictably.
Creating internal automations that remove repetitive work.
Prototyping new agent behaviors, testing them with real users, and deciding what’s worth scaling.
What we care about most
We care less about checklists and more about how you think and build.
You’re likely a good fit if you:
Enjoy breaking down messy, ambiguous problems.
Can move quickly without sacrificing judgment.
Think in systems, outcomes, and feedback loops.
Are comfortable wearing multiple hats: engineer, architect, product thinker.
Prefer shipping, learning, and iterating over polishing abstractions in isolation.
Experience & skills (signals, not gates)
You’ll likely thrive in this role if you:
Use Python as a thinking tool, not just a language - comfortable turning vague ideas into working systems and iterating quickly.
Enjoy tackling messy, ill-defined problems and shaping them into clear, testable solutions.
Have built LLMs in practice and understand their strengths, limits, and failure modes.
Are curious about agentic systems and workflows - you experiment, prototype, and refine rather than wait for “best practices.”
Think in end-to-end systems: inputs → reasoning → tools → outputs → feedback.
Care about outcomes and leverage, not just correctness or architectural purity.
Have a bias toward automation and eliminating manual work when software can do it better.
Are comfortable shipping imperfect versions, learning from real usage, and improving fast.
Naturally think about reliability, cost, and performance once something starts to matter.
Backgrounds that often work well:
Product-minded backend engineers.
Engineers who’ve acted as de-facto product owners.
Solution or systems architects who like to build.
People who’ve worked close to real users and production constraints.
Nice to have (not required)
Production monitoring and evaluation of AI systems.
Frontend awareness (you understand product flows even if you don’t build UIs).
Experience automating workflows across tools via APIs and webhooks.
Familiarity with agentic systems, RAG pipelines, or workflow orchestration.
Comfort working with modern LLM APIs (OpenAI, Anthropic, etc.).
Experience designing APIs and services (FastAPI / Flask / Django or similar).
Cloud experience (AWS / GCP / Azure).
Experience with LangChain, LangGraph, LlamaIndex, or similar frameworks
Ability to join occasional in-person sessions in Prague.
The way we work
Flat hierarchy, high ownership, low ego.
Fast iteration, honest feedback, strong trust.
We challenge ideas, including our own, and optimize for learning and impact.
We ship early, measure, and improve.
What you can expect
Fully remote work (EU-friendly).
Real, non-toy AI product challenges.
High autonomy and real influence on product direction.
Smart teammates who care deeply about what they build.
Unlimited vacation (as long as you’re learning and delivering).
How to apply
Send us something that shows how you think and build:
A résumé or short intro.
A GitHub repo, side project, or prototype.
A Loom demo or write-up of something you’ve built.
If you enjoy building AI systems that actually ship, break, improve, and matter, we should talk.