Experience
Google / Senior Research Scientist
San Francisco, USA | New York, USA / 2024 - Present
- Core contributor to Gemini Omni pre-training – co-designed multi-modal reference conditioning, built an agentic data-annotation harness, POC for an end-to-end vertical, run-captained large-scale experiments.
- Core contributor to Veo's reference-to-video capabilities – designed key conditioning components and data mixture critical to achieving state-of-the-art performance.
- Led the design of a novel model for grounding Genie in 3D; end-to-end from concept to working model.
- Published 15 top-venue papers – 1 Oral (top 0.4%), 5 Spotlights/Highlights (top 3%); mentored 15+ interns.
Google / Research Scientist
San Francisco, USA | London, UK / 2022 - 2024
- Led the first reference-to-video Veo model design and view-conditioned diffusion research, turning as few as one product photo into an interactive 3D experience – on Google Search for 1M+ products / billions of impressions.
- Built and scaled a distributed synthetic-data rendering pipeline that renders 1K frames per second.
Adobe Research / Research Intern
London, UK / 2021
- Researched generative texture & material modeling for 3D assets, advised by Milos Hasan and Kalyan Sunkavalli.
Meta AI / Research Intern
London, UK / 2020
- Researched few-shot 3D reconstruction from videos, advised by David Novotny and Andrea Vedaldi; code integrated into Meta's internal PyTorch3D.
- Published 2 papers, including an ICCV Best Paper Honorable Mention (top 0.25% of accepted papers).
Robert Bosch GmbH / Ulm University / Research Assistant
Ulm, GER / 2017 - 2018
- Co-authored two publications on object detection and tracking for autonomous driving.
Education
Ph.D. Computer Science at University College London (UCL)
2018 - 2023
No corrections · Eurographics Best PhD Thesis Award · Advisors: Tobias Ritschel and Niloy J. Mitra
M.Sc. Media Informatics at Ulm University
2015 - 2017
Grade 1.1 with distinction (~3.9 US GPA) · Advisor: Timo Ropinski
B.Sc. Media Informatics at Ulm University
2012 - 2015
Grade 1.3 (~3.7 US GPA) · Advisor: Enrico Rukzio
Product Impact
Genie Google DeepMind's real-time interactive world model.
2026
Gemini Omni A video model that creates videos from any input.
2026
Veo 2/3 Google DeepMind's state of the art generative video model with audio.
2025
Google Search 360-spins AI-generated 360-degree spins of products.
2022-2025
Awards & Honors
Eurographics Best PhD Thesis Award
2024
ICCV Best Paper Honourable Mention (top 0.25% of accepted papers)
2021
Rabin Ezra Scholarship
2021
Google Fellowship & Department of Computer Science UCL Scholarship
2018
1st Place, Audi App Challenge
2016
[15]
GR3EN: Generative Relighting for 3D Environments ·
SIGGRAPH, 2026
[12]
Bolt3D: Generating 3D Scenes in Seconds ·
ICCV, 2025
[9]
Can Generative Video Models Help Pose Estimation? ·
CVPR, 2025
[8]
IllumiNeRF: 3D Relighting without Inverse Rendering ·
NeurIPS, 2024
[7]
CAT3D: Create Anything in 3D with Multi-View Diffusion Models ·
NeurIPS, 2024
[5]
Generative Modelling of BRDF Textures from Flash Images ·
SIGGRAPH ASIA 2021
[3]
[2]
[1]
Single-Image Tomography: 3D Volumes from 2D Cranial X-Rays ·
Eurographics 2018
Mentoring, Teaching & Invited Talks
- Mentored 14 student researchers at Google and UCL; work published across top academic conferences.
- UCL Teaching Assistant – Computer Graphics (COMP0027), Machine Vision (COMP0137), Machine Learning (COMP0088).
- 15 invited talks at universities, research labs, and academic conferences; panelist at NYC Tech Week.
Academic Service
Area Chair: ICML, ICLR
Program Committee: Eurographics
Reviewer: CVPR, ICCV, ECCV, SIGGRAPH, SIGGRAPH Asia, EG, TPAMI, TCVG, NeurIPS, CGF