We are looking for a research intern to join us for a research project aimed at publication at a top-tier venue. The intern will design and develop novel systems that explore the interaction between human pose understanding and vision-language models (VLMs), advancing how these modalities can be combined to reason about human motion, activity, and embodied behavior across images and video.
Our group develops hand and body pose tracking algorithms for our various devices and applications. One such example includes the hand tracking input for the Vision Pro.
Design and implement novel methods that integrate pose representations with vision-language models, targeting established academic benchmarks.
Collaborate with researchers and engineers on the team to produce a publication-ready contribution.
Benchmark against established evaluation suites and iterate toward state-of-the-art results.
Requirements: Minimum Qualifications:
Currently enrolled in a graduate program (M.Sc. or Ph.D.) in Computer Science, Electrical Engineering, or a related field.
Publications at top-tier venues (e.g., NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, ACL, EMNLP or similar).
Strong programming skills in Python and experience with deep learning frameworks (e.g., PyTorch).
Solid foundation in computer vision, natural language processing, or multimodal learning.
Preferred Qualifications:
Demonstrated expertise working with Vision-Language Models (VLMs) and/or Large Language Models (LLMs).
Experience with human pose estimation, motion modeling, or related body-tracking tasks.
Familiarity with video understanding tasks and temporal modeling.
Familiarity with multimodal learning and benchmarks that combine language with visual or spatial data.
Experience with prompt engineering and optimization techniques.
This position is open to all candidates.