3D Machine Learning Engineer

About ALLSIDES ALLSIDES is a Deep-Tech startup from the European Alpine region with offices in Italy and New York, that secured significant funding in early 2025 (announced under our former name Covision Media) and is experiencing rapid growth. We're redefining how the world experiences 3D content. By combining physically accurate scanning and generative AI, we enable new content creation workflows for e-commerce, virtual environments, and immersive experiences. Our clients include global brands like adidas, Meta, Amazon, and Zalando. Using our own scanners, we've built one of the world's largest datasets of photorealistic, relightable 3D assets with accurate ground truth—capturing at production scale (tens of thousands annually) while simultaneously training next-generation AI models. Meta uses our scanning technology for their HOT3D dataset (ECCV 2024). We collaborate with Prof. Cremers at TU München, Prof. Rother at Heidelberg University's Computer Vision Lab, and are advised by Prof. Pietro Perona from Caltech. As a research-driven startup and NVIDIA Inception member, we operate at the forefront of 3D machine learning and visual computing. More info about us and our technology: Website: NVIDIA AI Blog: Position Overview We are looking for a 3D Machine Learning Engineer with deep expertise in neural rendering, inverse rendering, and generative 3D methods. You will develop learning-based systems that recover high-fidelity geometry and physically-based materials from captured imagery—pushing the boundaries on relighting, material decomposition, and generative reconstruction. This role sits at the intersection of deep learning and graphics. If you want to ship neural 3D systems at production scale while advancing the state of the art, this is for you. Key Responsibilities Development of Neural 3D Systems Build and optimize neural inverse rendering models for geometry and material estimation. Develop relighting and appearance decomposition pipelines using differentiable rendering techniques. Explore generative reconstruction approaches: diffusion-based 3D, feed-forward reconstruction networks, neural scene representations. Tackle challenging scenarios: specular and transparent surfaces, thin structures, complex BRDFs. Write clean, production-ready code and iterate on model quality and inference speed. Collaboration & Exploration Work with rendering specialists and system developers to integrate learned components into the pipeline. Engage with enterprise clients to gather requirements and iterate on capabilities. Stay current with CVPR, SIGGRAPH, and related research; identify methods to bring into production. Qualifications Required: Strong foundation in applying machine learning to 3D, gained through a PhD or Master's degree combined with 3+ years of industry experience. Proven track record in neural 3D: NeRF variants, Gaussian splatting, neural inverse rendering, or material decomposition. Experience with differentiable rendering frameworks (Mitsuba, nvdiffrast, PyTorch3D, or similar). Solid understanding of physically-based rendering fundamentals. Proficiency in Python and PyTorch. Demonstrated ability to take models from research to production. Nice to Have: Experience with generative 3D models (diffusion-based reconstruction, latent 3D representations). Background in classical computer vision: SfM, multi-view geometry, mesh processing. Experience with C++ in performance-critical pipelines. Familiarity with CUDA and GPU optimization. Experience with TensorRT, ONNX, or MLOps practices for model deployment. Published research or open-source contributions in neural rendering or 3D ML. What we offer Competitive salary package including stock options Amazing office space in the Italian Dolomites Personal and professional growth: own critical parts of our pipeline and help us scale Flexible working hours and work-from-home policy (1 day/week) Participate in CVPR, SIGGRAPH, and AI/startup conferences globally Lunch vouchers for working days Relocation assistance and visa sponsorship This is a full-time role located on-site at our Bressanone/Dolomites offices (South Tyrol, Italy). For exceptional candidates, part-time remote arrangements may be considered. Send your resume to careers@allsides.tech