About the role
Peepel is building the platform that simplifies people management for businesses. We connect HR, payroll, fleet, and workforce tools into one seamless experience — so companies can focus on their people, not their processes. We don't separate "product" from "engineering," and we don't treat AI as a side project. Our AI engineers own the full delivery lifecycle of intelligent features: they talk to customers, understand the workflows we can automate, design the systems, and ship them into production. You won't be wiring up demos — you'll be the Directly Responsible Individual (DRI) for AI features that real companies depend on every day. You care as much about what you build as how you build it. We ship fast, keep things simple, and lean on managed services wherever possible. You'll report to our CTO and work directly with customers, partners, and designers to decide where AI genuinely earns its place in the product — and where it doesn't.
Responsibilities
- Own AI features end-to-end — from spotting the customer pain through model choice, evals, shipping, and iteration - Talk to customers directly to find the workflows where automation, agents, or LLMs unlock real time savings - Shape requirements before writing code: prototype, measure, and define what "good output" looks like before committing to an architecture - Design and build agent systems, retrieval pipelines, and tool-using LLM workflows on top of our API and integrations - Build evals, guardrails, and observability so AI features stay reliable as models, prompts, and data evolve - Work with the "garage door up" — share prompts, traces, eval results, and trade-offs transparently - Keep cost, latency, and accuracy in balance as usage grows; pick the right model for the job, not the most impressive one - Mentor peer developers on what works (and what doesn't) when applying LLMs to production product work
Requirements
- We don't care how many years you've been doing this. We care what you've shipped. Show us LLM-powered features in production, agents you've built, evals you've run, prompts you've debugged at 2am. A side project that thousands of people use beats a decade of résumé bullets. - Fluent in TypeScript and Node.js — enough to be dangerous in a real codebase on day one - Hands-on experience with at least one major model provider (Anthropic, OpenAI, Google) and the good, bad and the ugly around it: tool use, structured outputs, streaming, prompt caching, evals - Strong data affinity — comfortable with relational databases, vector stores, retrieval pipelines, or integration architectures - You think in terms of customer problems and outcomes, not benchmarks. You know when an LLM is the right tool — and when a WHERE clause is. - Ability to break large AI initiatives into small, shippable, measurable increments — and kill the ones that don't pan out - Intellectually curious, independently driven, and passionate about great products. You read papers but ship code. - Pragmatic: you keep technical debt in check but know when to ship and iterate - Fluent in English and willing to work hybrid from our Ghent office
Benefits
- Impact: Early-stage company where every AI feature you ship has immediate, visible impact for real customers — not a research backlog. - Growth: Learn from seasoned entrepreneurs and grow fast as the company scales globally. - Culture: Collaborative team that values craftsmanship, pragmatism, and fun. - MacBook, healthcare plan, ESOP, meal vouchers, annual retreat, nice office, team events,…