Senior+ AI Researcher (Multimodal Perception Models)

Join Tavus to lead foundational research at the intersection of multimodal modeling and conversational AI. You'll drive the technical roadmap for systems that perceive, reason, and generate across video, audio, and language—shaping the next generation of real-time AI avatars that enable natural, face-to-face human-computer interaction.

What you'll do

• Lead foundational research on multimodal models for conversational avatars that perceive, reason, and generate across video, audio, and language.
• Build and train autoregressive, predictive, and diffusion-based architectures with deep focus on temporal and sequential data.
• Design and execute experiments to predict and control avatar responses across visual, auditory, and linguistic channels.
• Partner with the Applied ML team to translate cutting-edge research into real-world production systems.
• Mentor researchers and drive technical excellence across the team.

What Tavus is looking for

• 2–3+ years of hands-on experience with LLMs, VLMs, or multimodal systems, backed by a strong research background.
• Proven experience leading research efforts and mentoring teams.
• Deep expertise in sequence modeling across video, audio, and text, including autoregressive, predictive, and diffusion frameworks.
• Hands-on experience training and optimizing large-scale models for performance and real-time generation.
• Track record of translating research ideas into production-grade systems.
• Publications in top-tier venues such as CVPR, ICCV, NeurIPS, ECCV, or ACMMM.
• Strong PyTorch skills and ability to move fluidly between research and engineering.