Research Scientist Intern, 3d Scene Understanding (phd)

Meta · Big Tech · Zurich, Switzerland

Meta Reality Labs Research is seeking a PhD intern to conduct research in 3D scene understanding for AR/VR devices and embodied agents. The role involves research in areas like 3D representation learning, scene graphs, multimodal grounding, and object tracking, with a focus on advancing the research frontier and aiming for top-tier conference submissions. Experience with 3D datasets, deep learning frameworks, and hybrid 3D representations is required.

What you'd actually do

Conduct research in 3D scene understanding, including but not limited to: Efficient and scalable 3D representation learning or reconstruction, Hierarchical / semantic scene graphs for large-scale environments, Open-vocabulary multimodal grounding (text, image, video, audio etc), Dynamic object tracking and temporal scene reasoning, and Edge-aware 3D scene understanding.
Design and execute experiments on public datasets and/or internal data from AR/VR hardware.
Collaborate with researchers, engineers, and interns across Reality Labs to advance the research frontier.
Document findings and aim for submission to CVPR or equivalent top-tier conferences.

Skills

Required

Currently pursuing a PhD in Computer Vision, Machine Learning, Robotics, or a related field
Research background in at least one of: 3D scene understanding, semantic mapping, SLAM, or neural representations
Multimodal learning (vision + language, vision + audio)
Large-scale 3D reconstruction or scene graph generation
Programming skills in Python and PyTorch (or equivalent deep learning frameworks)
Hands-on experience with 3D datasets, rendering, or AR/VR sensors

Nice to have

Experience with hybrid 3D representations (e.g., Gaussian Splatting, NeRFs, octree/voxel methods)
Familiarity with open-vocabulary or foundation models (CLIP, Segment-Anything, LLMs for vision-language reasoning)
Knowledge of real-time or efficient deployment techniques (quantization, distillation, adaptive inference)
Ability to integrate research code into practical prototypes on edge devices (Jetson, AR/VR headsets)
Proven track record of achieving significant results as demonstrated by grants, fellowships, patents, as well as first-authored publications at leading workshops or conferences
Experience working and communicating cross functionally in a team environment

What the JD emphasized

Proven publication record in top-tier venues (CVPR, ICCV, ECCV, NeurIPS, ICLR, etc.)
Must obtain work authorization in country of employment at the time of hire, and maintain ongoing work authorization during employment
Intent to return to degree-program after the completion of the internship/co-op

Other signals

3D scene understanding
multimodal grounding
on-device intelligence
research frontier

Read full job description

Reality Labs Research is pioneering the next generation of AR/VR devices and intelligent systems. Our mission is to create the foundational technologies that enable contextual AI for AR glasses, VR headsets, and embodied agents. We are advancing the state of the art in 3D perception, multimodal scene understanding, and efficient on-device intelligence. We are seeking a highly self-motivated and hands-on PhD student to join our team as a Research Scientist Intern for Summer 2026, focusing on 3D scene understanding. Our internships are twelve (12) to twenty-four (24) weeks long and we have various start dates throughout the year.

Responsibilities

Conduct research in 3D scene understanding, including but not limited to: Efficient and scalable 3D representation learning or reconstruction, Hierarchical / semantic scene graphs for large-scale environments, Open-vocabulary multimodal grounding (text, image, video, audio etc), Dynamic object tracking and temporal scene reasoning, and Edge-aware 3D scene understanding. Design and execute experiments on public datasets and/or internal data from AR/VR hardware. Collaborate with researchers, engineers, and interns across Reality Labs to advance the research frontier. Document findings and aim for submission to CVPR or equivalent top-tier conferences.

Qualifications

Currently pursuing a PhD in Computer Vision, Machine Learning, Robotics, or a related field Research background in at least one of: 3D scene understanding, semantic mapping, SLAM, or neural representations, Multimodal learning (vision + language, vision + audio), Large-scale 3D reconstruction or scene graph generation Proven publication record in top-tier venues (CVPR, ICCV, ECCV, NeurIPS, ICLR, etc.) Programming skills in Python and PyTorch (or equivalent deep learning frameworks) Hands-on experience with 3D datasets, rendering, or AR/VR sensors Must obtain work authorization in country of employment at the time of hire, and maintain ongoing work authorization during employment Experience with hybrid 3D representations (e.g., Gaussian Splatting, NeRFs, octree/voxel methods) Familiarity with open-vocabulary or foundation models (CLIP, Segment-Anything, LLMs for vision-language reasoning) Knowledge of real-time or efficient deployment techniques (quantization, distillation, adaptive inference) Ability to integrate research code into practical prototypes on edge devices (Jetson, AR/VR headsets) Intent to return to degree-program after the completion of the internship/co-op Proven track record of achieving significant results as demonstrated by grants, fellowships, patents, as well as first-authored publications at leading workshops or conferences Experience working and communicating cross functionally in a team environment