Staff Research Engineer, Multimodal Generative AI

synthesia; United States; On-site; USD 180,000 - 260,000; Full time

Job summary

Synthesia is seeking a Staff Research Engineer to join the Voice team and drive a broader vision across teams. You will own design and implementation of core components for voice-to-voice models that synthesize text and speech in real time, collaborating with video teams and senior researchers.

Lead roadmap, prototype quickly, and ship end-to-end solutions from pretraining to production, while advancing multimodal interaction and latency-aware evaluation across products.

Qualifications

• Strong understanding of generative modelling applied to sequential or multimodal data.
• Hands-on experience with large language models or transformer-based architectures.
• High proficiency in PyTorch, including distributed training and model optimisation.
• Solid grasp of time-series modelling and tokenisation, preferably in audio, speech, or video.
• Proven ability to prototype quickly, test hypotheses, and iterate efficiently.
• Experience training deep learning models end-to-end, from data preparation through evaluation.
• Strong general software engineering skills for a large research infrastructure.

Responsibilities

• Shape roadmap to create new model capabilities and unlock functionality for customers.
• Propose novel multi-modal architectures (text and voice).
• Develop and evaluate streaming and conversational systems for low-latency synthesis.
• Design solutions that reinforce emotional expressiveness and natural interaction.
• Implement from pretraining through post-training and shipping to production.
• Integrate and test architectures (neural codecs, diffusion, flow-matching) to enhance realism.
• Define new evaluation metrics for conversational systems including latency-aware measures.
• Track the latest research in audio-visual diffusion, autoregressive models, and multimodal LLMs.
• Curate new datasets to complement existing data.
• Lead post-training initiatives like DPO, fine-tuning, and distillation to shipping quality.
• Ship models to production with optimised runtime and address feedback.

Skills

Generative modelling

Transformer architectures

PyTorch distributed training

Time-series modelling

Rapid prototyping

End-to-end training

Software engineering

Job description

Synthesia is seeking a Staff Research Engineer to join the Voice team and drive a broader vision across teams. You will own design and implementation of core components for voice-to-voice models that synthesize text and speech in real time, collaborating with video teams and senior researchers.

Lead roadmap, prototype quickly, and ship end-to-end solutions from pretraining to production, while advancing multimodal interaction and latency-aware evaluation across products.