A company in our network is looking for an engineer who builds AI-powered products, not just models. Youre fluent in the modern LLM toolkit — prompting, retrieval, tool use, agentic workflows — but you reach for those tools in service of a user outcome, not the other way around. You care about what lands in front of people and how it feels to use.

Day to Day

Ship user-facing AI products. Take ideas from prototype to production and get them in front of real users. Youll work across the stack when needed to close the gap between whats technically possible and what actually feels good to use.

Own the user experience. Drive how AI features behave from the users point of view — the interaction, the response quality, the failure modes, the moments where it either delights or frustrates. You treat UX as a first-class engineering concern, not something handed off to design at the end.

Build the AI systems behind the product. Design and implement RAG pipelines, prompt management, and agentic workflows that hold up under real usage. You know these techniques well enough to choose the right one for the job and to know when a simpler approach wins.

Make pragmatic model decisions. Decide when to prompt, when to reach for retrieval, when to fine-tune, and when the answer isnt a model at all. You explain those tradeoffs clearly to engineers and non-engineers alike.

Keep it honest and reliable. Instrument and evaluate outputs so you catch hallucinations and regressions early. Build guardrails that survive contact with real users and real load, and keep an eye on latency, cost, and reliability once things are live.

Partner on the problem, not just the solution. Work closely with product and design to frame problems well before implementation starts.



Push back when the framing is wrong, and help the team stay focused on whats actually worth building.

Bring the landscape back to the team. Stay current on tooling and techniques — orchestration frameworks, emerging agent patterns, evaluation approaches — and translate the relevant advances into things the team can actually use.

Who You Are

Beyond the qualifications, we hire through a specific lens. These arent buzzwords; theyre what well actually look for in how you talk about your work.

You lead with the user. You instinctively ask "who is this for and what will it feel like to use" before "whats the cleanest architecture." You can hold both, but the user comes first. You have real product judgment — you can tell when something shouldnt be built, and you say so.

Youre a builder, not a maintainer. Youre most energized when there isnt a clear path yet and you get to define it. You identify gaps, shape solutions, and drive them forward without waiting for direction. You operate as much like a strategic partner as an implementer — influencing decisions and guiding stakeholders, not just taking tickets.

Youre genuinely AI-curious. Youre not using AI tools occasionally; youve rewired how you work around them. Youre faster, sharper, and more prolific because of it, and you bring that energy to how you build, prep, communicate, and think. You see AI as a multiplier,



not a shortcut.

Must-Haves

• Hands-on experience building and shipping LLM-powered product features, with working command of RAG, prompt management, and tool use.
• Strong product judgment: the ability to evaluate whether something should be built before figuring out how, and to reason about user experience end to end.
• Strong full-stack fundamentals and genuine comfort moving across backend and frontend to ship complete, user-facing products.
• A track record of taking AI features from prototype to production, including basic evaluation, monitoring, and attention to latency, cost, and reliability in live environments.
• Exceptional communication skills — you can explain complex technical decisions clearly to engineers, product managers, and executives.
• Experience owning outcomes across team boundaries: spotting capability gaps, aligning engineering and product, and influencing how a broader group approaches AI.

Nice-to-Haves

• Experience building agentic or multi-step reasoning systems, whether with tools like LangChain and LlamaIndex or a custom orchestration framework.
• A clear point of view on when fine-tuning outperforms prompting, ideally from having done both.
• Solid understanding of modern model architectures (transformers, diffusion models) and informed opinions on when to apply them.
• Familiarity with AI safety considerations, guardrail frameworks, and responsible deployment.
• Background in a SaaS or product analytics environment where user behavior data informs AI product design.
• Prior experience contributing to or launching a net-current team or product area.