cv
Basics
| Name | Jack Saunders |
| Label | Senior Research Scientist | AI Avatars & Digital Humans |
| jacksaunders909@gmail.com | |
| Url | https://jsaunders909.github.io/ |
| Summary | Senior Research Scientist at Tavus working on next-generation avatar rendering and generation. PhD from the University of Bath (ART-AI CDT & CAMERA). Founder of RealSync AI. Previously interned at Microsoft and Epic Games. Research focuses on video diffusion models, Gaussian Splatting, and audio-visual multimodal AI for photorealistic digital humans. |
Work
- 2025.03 - Present
Senior Research Scientist
Tavus
Working on next-generation avatar rendering and generation technology, with a focus on video diffusion models for photorealistic, real-time avatars.
- Video diffusion models for photorealistic avatar generation
- Autoregressive generation methods (Self-Forcing) for real-time deployment
- Real-time avatar rendering and generation pipelines
- Gaussian Splatting and neural rendering
- 2024.06 - 2024.09
Research Science Intern
Microsoft
Research internship in the Mixed Reality & AI Lab. Led development of GASP (Gaussian Avatars with Synthetic Priors), a method for creating animatable 360-degree Gaussian avatars from a single image, published at CVPR 2025. Also contributed to RwR (Retrieve while Reasoning).
- Led development of GASP: Gaussian Avatars with Synthetic Priors (CVPR 2025)
- Real-time 360-degree animatable avatars from a single image at 70fps
- Contributed to RwR: Retrieve while Reasoning (CVPR 2025)
- Gaussian Splatting, synthetic data priors, avatar fitting
- 2024.04 - Present
Founder
RealSync AI
Founded RealSync AI to provide consultancy and science communication for avatar and digital human companies. Focus on lip-sync, talking face, and deepfake detection technologies.
- Consultancy for avatar and digital human companies
- Science communication around AI-generated video
- Lip-sync and talking face technology
- 2022.05 - 2024.05
Research Scientist (Part-time)
DeepReel
Employee #1 at DeepReel. Designed and implemented the core AI Avatar technology powering the product.
- Designed and built core AI avatar pipeline from scratch
- Lip-sync and talking face generation
- Neural rendering for video avatars
- 2021.11 - 2022.05
Research Intern
Epic Games
Research internship focused on style transfer for 3D facial animation, which later became the basis for the FACTS paper at Eurographics 2024.
- Style transfer for 3D facial animation
- Unreal Engine MetaHuman integration
- GAN-based emotion and style transfer
Education
Awards
- 2019.06.01
E T Davis Prize for Pure Mathematics
University of Southampton
Awarded to the student with the highest mark in Pure Mathematics in the final year of the BSc Mathematics course.
Publications
-
2025.06.01 GASP: Gaussian Avatars with Synthetic Priors
CVPR 2025
Real-time animatable Gaussian Splatting avatars from a single image or monocular video, using synthetic priors. Renders at 70fps on commercial hardware.
-
2025.01.01 DEAD: Data-Efficient Audiovisual Dubbing using Neural Rendering Priors
BMVC 2025
High-quality video dubbing from seconds of training data using a multi-person neural rendering prior.
-
2024.11.01 TalkLoRA: Low-Rank Adaptation for Speech-Driven Animation
BMVC 2024
Personalised speech-driven facial animation using LoRA adaptors with efficient transformer inference via chunking.
-
2024.04.01 FACTS: Facial Animation Creation using the Transfer of Styles
Eurographics 2024
Style transfer for 3D facial animation — retargeting emotion and idiosyncratic speaking style while preserving lip sync.
-
2023.11.01 READ Avatars: Realistic Emotion-controllable Audio Driven Avatars
BMVC 2023
Emotion-controllable audio-driven avatars using neural texture models conditioned on emotion with adversarial training.
Skills
| Deep Learning for Digital Humans | |
| Video Diffusion Models | |
| Gaussian Splatting | |
| Flow Matching | |
| GANs | |
| Transformers | |
| LoRA | |
| Neural Rendering |
| Audio-Visual AI | |
| Lip-sync | |
| Talking Face | |
| Speech-Driven Animation | |
| Audio-Visual Models | |
| Video Dubbing |
| Python | |
| PyTorch | |
| CUDA | |
| Hugging Face | |
| Diffusers |
Languages
| English | |
| Native Speaker |
Interests
| Digital Humans & Avatars | |
| Real-time rendering | |
| Video diffusion | |
| Self-Forcing | |
| Photorealism |