Principal ML Engineer — AR & Generative Visual AI
5 hours ago
Valencia
Sklum customers already design rooms in 3D through HomeByMe, a 3D product experience that runs across our site today. We're rebuilding it into something dramatically richer — AR room design with real Sklum products, AI-generated room imagery, in-context product visualization, virtual staging, and scene composition. It's the centerpiece of our Digital & AI vision and the single largest customer-experience bet Sklum is making over the next two years. We're hiring the person who owns the technical build end to end. Vendors don't build exactly what a retailer needs out of the box — closing that gap takes someone who can push a vendor roadmap, not just consume it. Open-source models move faster than any internal roadmap, which means the right stack today won't be the right stack in six months. And most "visual AI" hires split into two types who won't get this done alone: researchers who won't ship, or integrators who won't push the underlying technology far enough. We need someone who does both — decides the right stack across vendor, open-source, and in-house build, and personally writes the production code that ships it. No committee, no legacy constraints. The calls on architecture, vendors, and models are yours, and you carry them from decision to production. • The build. HomeByMe 2.0 end to end — AR, 2D, and 3D room and product visualization, built alongside our Principal UX Product Manager and design team from day one, not handed to you after design is locked, • Vendor strategy. Own the technical relationship with our existing HomeByMe vendor and evaluate new ones as the space evolves. Push roadmaps, surface real problems, decide where to deepen a partnership and where to build in-house instead. Every build-vs-buy-vs-steer call gets defended in writing, • The stack. Combine the best commercial and open-source AR, 2D, and 3D AI models — image generation, scene understanding, depth and 3D reconstruction, on-device inference — into one coherent system, then build Sklum's own layer on top: room understanding, product placement logic, scene composition, and the experience design that makes it unmistakably ours, • Production reliability. Ship code that runs at consumer scale, on real phones, in real network conditions, across every market we serve. Own inference infrastructure, on-device deployment, model serving, latency, cost, and the observability that tells the team when something is working — and when it isn't, • Cross-functional connection. Sit at the center of science, product, UX design, engineering, content, brand, and merchandising. Partner directly with the Principal UX Product Manager, the UX tech team on AR integration into our app, and merchandising on the product catalog data that visual AI depends on, • Technical leadership. Set the bar for what production ML looks like at Sklum — code quality, evaluation discipline, deployment rigor — and lead through influence and technical standard, regardless of formal reporting lines, • Agentic engineering. Personally use coding agents and autonomous workflows as a core part of how you build, evaluate, deploy, and iterate. Design the human-in-the-loop oversight and evaluation harness that let agentic workflows safely contribute to production ML code, and keep raising the team's leverage as the tooling improves, • 10+ years of engineering experience, with a strong and recent focus on production machine learning — not historical, active. Candidates with meaningfully more experience are welcome, • Principal IC track record — you have personally written production code and shipped customer-facing ML products at scale. We're not hiring a manager who used to ship, • Deep, current understanding of modern machine learning, deep learning, and generative AI — model families, training paradigms, evaluation techniques, and the practical engineering around them. You read papers when they matter, but you ship products, • Strong engineering fundamentals — code review, testing, observability, infrastructure-as-code, and CI/CD as second nature, not a checklist, • Demonstrated experience building and shipping production ML systems at consumer scale — not internal tools, not demos, not research projects. Real systems running in front of real customers, • Meaningful depth in at least one of: AR, computer vision, 3D reconstruction and generation, scene understanding, generative imaging, or modern image/video diffusion models — you don't need to have built an AR room designer before, but you need the foundation to learn fast across the 2D, 3D, and AR landscape this role spans, • Track record of integrating commercial and open-source ML solutions into one coherent stack, with build-vs-buy decisions you've made and can defend, • Demonstrated experience working with — and steering — major enterprise software vendors: pushing a roadmap, escalating when delivery slips, translating product vision into vendor-shaped requests, • Strong opinions on the difference between a research prototype and a production system, and a track record of bridging the two, • Demonstrated personal mastery of agentic ways of working — you use coding agents and autonomous workflows as a core part of your own engineering practice, not as an experiment, • Track record of close partnership with product managers and designers on products where the model and the experience are designed together, not handed off, • Comfortable operating at the center of a broad cross-functional team — product, design, engineering, content, brand, and merchandising, all in the same week, • Rapid iteration mindset — comfortable in environments that ship and learn on a weekly rhythm, not a quarterly one, • Proven ability to lead effectively across cultures, locations, and time zones in a distributed, async-first environment, • Strong written and verbal communication — able to explain ML trade-offs clearly to non-ML audiences. Fluent English, • On-device ML and mobile inference experience — especially relevant given the AR component of this role, • Background in consumer-facing ecommerce, marketplace, lifestyle, home, or design-led products, • Experience with 3D/CAD/spatial computing platforms — Autodesk, Unity, Unreal, or similar, • Working proficiency in Spanish, • Experience explaining technical trade-offs to non-technical stakeholders in fast-moving, ambiguous environments your primary output is research papers, methodology, or theoretical work rather than shipped product. We're not looking for a pure research scientist, and we're not looking for a statistics-led data scientist whose tools are notebooks and analyses. If you read papers to make sharper engineering decisions and then ship, this is exactly right. Within your first year, HomeByMe 2.0 is live for real customers — AR room design, AI-generated imagery, and in-context product visualization running reliably at scale, not in a lab. You've made and defended real build-vs-buy calls across at least one major vendor relationship and one open-source integration. Agentic workflows are a visible, adopted part of how the ML function ships code, not a side experiment. And the partnership between science, product, design, and merchandising runs without you being a bottleneck in the middle of it. Based in Valencia, Spain, with a hybrid working policy. Relocation support is available for candidates moving from elsewhere. Benefits include health insurance, a professional development budget, childcare support, employee discounts, and gym/public transport discounts. Sklum is PE-backed, profitable, and scaling — what you build here goes live at real consumer scale, not shelved after the next funding round. Sklum Digital & AI is an agentic-native team: agentic AI workflows aren't a productivity layer on top of how we work, they're how the work gets done. You'll join a small, lean team with the autonomy of a principal IC and the backing of a company that's already serving customers across Europe today. #J-18808-Ljbffr
Waiters
Sales & Marketing
Chef & Cook
Retail
Salon & Beauty
Construction & Trades
Accounting & Finance
Animal Care
Art, Media, Design
Bar Staff
Barista
Child Care
Cleaning
Driver & Delivery
Education
Engineering
Entertainment
Events & Promotion
Fashion
Health & Social Care
Host & Hostess
Information Technology
Kitchen Porter
Legal
Management
Manufacturing
Office & Admin
Science
Security
Sports & Wellness
Warehouse
Writing & Editing
Other
