Staff Research Scientist- Interactive Avatars
Synthesia · Remote · Europe · London · Munich · Zurich ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Staff level research role
- Focus on video diffusion
- Hybrid or remote (Europe)
- Mentorship responsibilities
- Generative AI focus
Synthesia seeks a Staff Research Scientist to lead research in dyadic interaction modeling and avatar-centric video diffusion. You will develop models that understand user audio and video inputs to generate contextually appropriate, real-time reactions including gaze and body language. This role involves setting research roadmaps, mentoring junior researchers, and translating breakthroughs into product features.
What you'll do
- Define research direction for dyadic interaction modeling
- Advance perceptual layers for interactive agents
- Post-train multimodal models for natural interactions
- Adapt diffusion models to conversational signals
- Build evaluation frameworks for interaction quality
- Collaborate with data teams on dataset creation
- Mentor researchers and drive technical decisions
What you'll need
- Deep expertise in diffusion models for video/avatars
- Strong publication record in top-tier ML venues
- Experience leading and mentoring research teams
- Track record of moving research to production
- Proficiency in PyTorch and large-scale training
- Ability to communicate hypotheses and results clearly
Nice to have
- Experience with real-time streaming generation
- Knowledge of distillation for low-latency inference
- Background in audio-driven motion modeling
- Expertise in conversational modeling and turn-taking
Stack
- PyTorch
- Diffusion Models
- Multimodal Models
How you'll work
- Hybrid work setting available
- Remote options within Europe
- Office hubs in London, Zurich, Munich
What you get
- Competitive compensation package
- 25 days annual leave plus public holidays
- Regular planning and social events
- Location-dependent additional benefits