Staff Research Engineer – Multimodal Generative Modeling

About the role

Synthesia is building interactive multimodal systems that combine text, audio, and video in real‑time. As a Staff Research Engineer you will help shape the roadmap, propose ambitious research directions, and deliver critical components that power next‑generation voice‑video experiences.

Key responsibilities

• Shape the roadmap to create new model capabilities and unlock functionality for customers.
• Propose novel multimodal system architectures, especially for text and voice.
• Develop and evaluate streaming, low‑latency conversational systems for voice‑video synthesis.
• Design solutions that reinforce emotional expressiveness and natural interaction.
• Implement designs from pre‑training through post‑training and ship models to production.
• Integrate and test novel architectures such as neural codecs, diffusion, and flow‑matching models.
• Define new evaluation metrics for conversational systems, including latency‑aware measurements.

Required profile

• Ability to bring novel ideas and designs that advance interactive multimodal systems.
• Strong understanding of generative modelling applied to sequential or multimodal data.
• Hands‑on experience with large language models or similar transformer‑based architectures.
• High proficiency in PyTorch, including distributed training and model optimization.
• Solid grasp of time‑series modeling and tokenization, preferably for audio, speech, or video.
• Proven experience training deep learning models end‑to‑end, from data preparation through evaluation.
• Strong general software‑engineering skills for contributions to a shared research infrastructure.

Required skills

• PyTorch
• Distributed training
• Model optimization
• Large language models (transformer‑based)
• Time‑series modeling
• Audio/speech/video tokenization
• Deep learning model training