Research Scientist- Interactive Avatars

This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a Staff Research Scientist – Interactive Avatars based in Germany.

This is a senior research role at the forefront of generative AI, focused on building the next generation of interactive, avatar-based video agents. You’ll work within a multidisciplinary R&D organization combining research, engineering, and data expertise. Your focus will be on creating models that can understand audio and visual signals and respond with natural, human-like conversational behavior. You’ll shape research direction while balancing ambitious long-term opportunities with near-term product impact.

The role combines deep technical research, hands-on model development, evaluation, and research leadership. You’ll also mentor researchers and influence technical decisions across multiple teams, helping transform breakthrough ideas into production-ready capabilities.

• Define the research direction and roadmap for dyadic interaction modeling, balancing long-term research opportunities with immediate product priorities.
• Advance the perceptual capabilities of interactive AI agents, including understanding user audio and video and generating contextually appropriate responses.
• Develop and post-train multimodal models capable of producing rich, natural, and responsive dyadic interactions from audio and video inputs.
• Adapt and extend diffusion models to incorporate conversational conditioning signals such as dialogue state, turn‑taking, listener cues, and contextual reactions.
• Research and develop models capable of generating natural conversational behaviors, including gaze, facial expressions, body reactions, and back channel responses.
• Design robust evaluation frameworks, benchmarks, and automated test suites to continuously measure interaction quality.
• Collaborate closely with data specialists to define data requirements and shape high‑quality datasets for model development and training.
• Lead technical decisions spanning research, data, and engineering, ensuring strong alignment between research goals and product requirements.
• Lead and mentor a small team of researchers, supporting their technical development and helping establish a high‑performing research culture.
• Drive research projects from initial hypothesis and experimentation through validation, productionization, and product impact.
• Communicate research hypotheses, methodologies, experimental results, and technical recommendations clearly to both research and cross‑functional audiences.

• Deep expertise in machine learning with extensive hands‑on experience working with diffusion models, ideally applied to video generation, avatar generation, or related multimodal applications.
• Strong research track record demonstrated through publications at leading conferences such as CVPR, ICCV, ECCV, NeurIPS, ICML, ICLR, or SIGGRAPH, or equivalent evidence of significant research impact.
• Expertise in areas such as video diffusion, world models, dyadic interaction modeling, multimodal generation, or related fields.
• Proven experience leading a small team of researchers and mentoring junior research talent.
• Demonstrated ability to take research concepts from initial idea through experimentation and into production systems.
• Strong proficiency in PyTorch and modern machine learning tooling for large‑scale model training.
• Strong understanding of multimodal learning, generative modeling, and large‑scale ML experimentation.
• Excellent analytical and problem‑solving abilities, with a rigorous approach to designing experiments and evaluating results.
• Clear communication skills and the ability to explain complex technical concepts, influence…