Google DeepMind · London
Research Engineer, Multimodal and Video Modeling
12/2/2025
Description
We’re looking for a Research Engineer with exceptional programming and engineering skills, a deep understanding of large-scale neural network training and data processing, and a strong working knowledge of machine learning experimentation. The ideal candidate will have hands-on experience with training and developing multimodal models.
Key responsibilities:
- Develop, maintain, and improve large-scale multimodal models, with a focus on video modeling, and the data pipelines for training and evaluation.
- Design and implement novel methods for evaluating and improving multi-modal generative models, particularly at the post-training stage.
- Build and maintain robust data pipelines for collecting and processing large-scale datasets, including human-labeled data.
- Collaborate with research scientists to translate research ideas into production-ready code.
Qualifications
- BSc, MSc or PhD/DPhil degree in computer science, physics, mathematics, applied statistics, machine learning, or equivalent practical experience.
- Strong background in deep learning, with proven experience with relevant architectures (e.g., Transformers, GNNs, etc).
- Experience with training large-scale machine learning models, particularly in the multimodal domain (video, text).
- Excellent software engineering skills with a proven ability to build robust and scalable systems.
- Proficiency with ML and scientific libraries such as JAX, TensorFlow, PyTorch, NumPy, and Pandas.
- Experience with either large-scale data processing frameworks (e.g., Apache Beam, Spark) or distributed training infrastructure.
- Experience with LLMs, VLMs, and/or diffusion models training and inference.
- Experience with Computer Vision tasks (video classification, object tracking, depth estimation etc.)
- Experience working in a team on projects from the initial proof-of-concept stage to final implementation and evaluation.