2027 Intern - Research Scientist

Internship Details

Location: London

Duration: 3 months

Starting date: From February 2027

End date: No later than 30 November 2027

Eligibility: master or Ph.D. student

Hybrid: 3 days per week in the office, 2 days per week work from home

Annual Pay: £75,000

About the role

Research at the intersection of 3D, Image, Video, and Generative AI

Can we bridge the compact, physically consistent structure of 3D representations with the rich temporal priors of generative video models?

Join Adobe Research London to tackle core problems in multi-modal generation and controllable editing. We are looking for PhD and MS researchers to explore how unifying image, video, and 3D modalities can yield richer, more controllable generative spaces.

The Research Focus

• Bridge 3D & Generative Video: Investigate compact 3D representations (e.g., neural fields, Gaussian splatting, explicit geometry) alongside video diffusion/autoregressive models to unlock precise spatial and temporal control.
• Combine Generative Models & Classical Rendering: Explore hybrid pipelines integrating physically based rendering (PBR) with generative/agentic workflows.
• Top-Tier Publication: Develop novel methods and target impactful publications at top-tier venues (SIGGRAPH, ACM TOG, CVPR, ICCV, NeurIPS, ICLR).
• Thesis Alignment: We actively prioritize research problems that can serve as core chapter(s) of your dissertation and establish sustained collaborations well beyond the internship.

Resources & Environment

• Compute & Infrastructure: Large-scale GPU clusters and specialized datasets to train and evaluate state-of-the-art models.
• Mentorship & Impact: 1:1 mentorship from senior Adobe research scientists, with a direct pipeline to transfer groundbreaking ideas to creative tools used by millions worldwide.

Qualifications

• Currently pursuing a PhD or MS in Computer Science, specializing in Computer Vision, Graphics, Machine Learning, or a closely related field. Exceptional senior undergrads with strong research output will be considered.
• Solid track record or coursework in generative modeling (diffusion / flow matching / autoregressive), computer graphics / vision, and/or 3D scene representation.
• Proficient in modern deep learning frameworks (PyTorch) with strong experimental rigor.

What to expect from the recruitment process

Our selection process consists of two stages as follows

• 60 minutes Technical Interview with the team.
• 45 minutes Soft-Skills Interview with the Hiring Manager.