AI Engineer

LEGO Digital Play (LDP) will bring the LEGO brand into digital entertainment in new, innovative, and wholly-owned ways. Our mission is to ensure the LEGO brand remains as powerfully a part of children's lives in the coming decades as it has ever been. We aim to reach every kid on the planet, their parents, and adult fans of LEGO—and provide them with meaningful, magical, and playful new experiences.

We are at the earliest phases of this new company, offering a unique opportunity to build a new entity for the world's most beloved and trusted brand. Our culture is open, collaborative, intellectually rigorous, and creatively vibrant.

Job Summary
As an AI Engineer, you will design, train, and deploy the core models that enable LEGO Digital Play's AI to generate with bricks in 3D. From brick recognition and graph-based model approaches to assembly planning and generative building, you'll transform ambitious research into production-ready systems.

This is a hands-on role at the heart of LDP's AI pillars. You'll push the boundaries of computer vision, 3D geometry, and generative modelling—scaling models into services that power game engines, SDKs, and hybrid play experiences. Working closely with product, data, and back-end engineers, you'll own performance, evaluation, and optimization, ensuring LEGO Digital Play's AI is safe, brand-aligned, and ready to unlock new worlds of play

Key Responsibilities

• Prototype, train, and productionize computer vision and 3D models, including segmentation, pose/part detection, connection-graph prediction, and GNN-based approaches.
• Develop generative models (diffusion, transformers, or similar) for text-to-build workflows, creative model manipulation, and LEGO brand specific 3D content generation.
• Design and run rigorous benchmarks, data augmentation pipelines, and active learning loops to improve accuracy and robustness over time.
• Optimize inference performance through quantization, distillation, CUDA kernels, and other techniques to meet strict latency and throughput SLOs.
• Ship scalable APIs and services in partnership with back-end engineers; own experiment tracking, model registry, and versioning.
• Collaborate with visual experience engineers and technical artists to validate and refine model outputs, raising the visual fidelity and play quality bar.
• Partner with gameplay and product engineers to integrate AI outputs seamlessly into interactive play loops.
• Contribute to responsible AI practices by embedding safety, fairness, and brand alignment into every stage of model development and deployment.

Qualifications

• Experience in applied ML experience with strong skills in PyTorch or TensorFlow, Python, and scientific tooling.
• Expertise in 3D machine learning: point clouds, meshes, voxel grids, geometric deep learning, GNNs, diffusion, and transformers.
• Experience of computer vision at scale, including dataset design, evaluation frameworks, error analysis, and mitigation of bias/safety issues.
• Experience deploying and optimizing models for production use (e.g., Triton/TorchServe, ONNX, TensorRT), GPU profiling, and performance tuning.
• Familiarity with quantization, pruning, distillation, and CUDA kernels to meet tight latency/throughput SLOs.
• Prototype-first and evidence-driven mindset: comfortable with rapid iteration, ambiguity, and delivering working proofs-of-concept.
• Strong scientific rigor: able to design clean experiments, ablations, and reproducible pipelines.
• Advocate for responsible AI, ensuring outputs are aligned with LEGO's values, brand/IP integrity, and child-safety requirements.
• Collaborative and cross-functional: works effectively with data engineers, infra engineers, gameplay engineers, and technical artists to ship production-quality features.

Preferred Qualifications

• Experience with physics or simulation frameworks (e.g., Havok, Bullet, or equivalent).
• Knowledge of engine integrations (Unity/Unreal) to visualize and test AI outputs in interactive contexts.
• Proficiency in C++ for high-performance model deployment or engine-side optimizations.
• Exposure to generative modeling for 3D content (e.g., diffusion, text-to-3D, procedural geometry).
• Interest in hybrid play and an understanding of how real-time AI enhances creativity and interactivity.