Staff Research Scientist: Interactive Avatar AI

Synthesia; Greater London; Hybrid; GBP 120,000 - 180,000; Full time

Benefits offered by this job

Hybrid work setting

25 days annual leave

Office in London

Job summary

Synthesia is seeking a Staff Research Scientist to join our Interactive Avatars team within the R&D department, focusing on avatar-centric diffusion models for interactive video. You will shape models that listen, respond, and react in real time, turning breakthroughs into product capabilities shipped to users.

You’ll lead a small team, mentor researchers, and collaborate with data and engineering to translate research into production features for millions of users worldwide.

Qualifications

• Deep ML expertise with diffusion models for video/avatar generation.
• Strong publication record at top venues (CVPR/ICCV/ECCV/NeurIPS/ICML/ICLR/SIGGRAPH) or demonstrated impact.
• Experience leading a small team of researchers and mentoring juniors.
• Track record of taking research from idea to production.
• Proficiency in PyTorch and modern ML tooling for large-scale training.
• Ability to communicate hypotheses, experiments, and results clearly.

Responsibilities

• Set research direction and roadmap for dyadic interaction modeling.
• Advance perceptual capabilities for interactive agents (audio/video understanding).
• Post-train multimodal models to generate rich dyadic interactions.
• Adapt diffusion models to new conditioning signals (conversational state, turn-taking).
• Build evaluation frameworks and test suites for interaction quality.
• Partner with data team to define data needs and build high-quality datasets.
• Lead and mentor researchers, driving decisions across research, data and engineering.

Skills

Diffusion models

PyTorch

Large-scale ML tooling

Clear communication

Tools

PyTorch

Job description

Synthesia is seeking a Staff Research Scientist to join our Interactive Avatars team within the R&D department, focusing on avatar-centric diffusion models for interactive video. You will shape models that listen, respond, and react in real time, turning breakthroughs into product capabilities shipped to users.

You’ll lead a small team, mentor researchers, and collaborate with data and engineering to translate research into production features for millions of users worldwide.