AI engineer in generative imaging

грейд
senior
формат
гибрид Барселона
страна
Испания
язык
Английский B1
опубликована
вчера

описание

Работа из офиса в Барселоне один день в неделю; необходимо иметь право на работу в Испании.

We are building the operating system for fashion design: an agentic platform that turns brand DNA into production-ready visuals, tech packs and collections in minutes instead of weeks. Our customers are global brands and retailers whose designers, merchandisers and suppliers depend on the platform to ship their next collection. We are a small, technical, fashion-native team. We move fast because our customers do. You will own the technical direction of our AI platform: the agents that power the product, the infrastructure behind them, and the layer of model adaptation that makes our outputs specific to each brand.

задачи

  • Design and build the AI agents at the core of our platform: multi-step workflows, tool use, MCP integrations, and the orchestration patterns that make them reliable and scalable in production.
  • Decide where customisation beats prompting and take it to production: adapting generative models to brand-specific data, training pipelines, versioning and rollback.
  • Design and maintain the systems that give our agents access to brand knowledge: embeddings, vector retrieval and the tradeoffs between accuracy, speed and storage.
  • Know which capabilities to build in-house and which to buy, and make those decisions based on quality, latency and cost, not habit.
  • Define how we know the system is working: automated judges, human review loops, regression benchmarks and the infrastructure to run them continuously.
  • Set the engineering standards for AI work across the team.
  • Write design docs, review code, mentor engineers and be the person the team turns to when a problem is hard.

требования

  • 6+ years building ML or AI systems, with a track record of shipping to real users at scale.
  • Deep experience designing and operating multi-step AI agents: tool use, orchestration, failure handling and evaluation.
  • You have fine-tuned or adapted generative models in production, not only in notebooks.
  • Strong Python and production engineering: tests, CI, observability and cost awareness.
  • Experience with embeddings, retrieval and semantic search at production scale.
  • You design evaluations that others trust, and you can tell a real improvement from noise.
  • You lead through technical judgement and clear writing: design docs, decision records and code reviews.
  • Daily use of AI-assisted coding tools.
  • Able to work from our Barcelona office one day a week, with the right to work in Spain.
  • Будет плюсом: Experience with model optimisation for inference: quantisation, distillation, efficient serving; Publications or open-source contributions in generative AI or agent systems; Experience working with creative or brand data, including licensing and consent; Interest in fashion, design tooling or creative software

условия

  • Competitive salary, calibrated to experience. We share the salary range with every candidate before the first interview.
  • ESOP (phantom shares): participation in our employee stock option programme, with an allocation sized to an early, high-impact role.
  • Compute and models: the budget to use the right model for the job, not the cheapest one, and access to real production data from real fashion brands.
  • You shape the direction.
  • We hire staff engineers to set the technical bar, not to follow a plan written by someone else. If you see a better approach, propose it, make the case and own the result.
  • Evidence over opinion.
  • Every change, whether it’s a model, a training decision or a system design, ships with the numbers that show it made things better. We trust data more than intuition, and intuition more than consensus.
  • Customers first.
  • The measure of any technical decision is whether it makes the product better for the designers using it. We stay close to real feedback and let it drive priorities.
  • Ownership, not tickets.
  • You own whole systems. You set the standards, decide what “done” looks like and are accountable for what runs in production.
  • Decisions are written down.
  • Short design docs and decision records mean the team doesn’t have to rediscover why something was built the way it was. Clear writing is a form of technical leadership.

навыки

компания

похожие вакансии

статьи для разработчиков