Multi-modal Representation Learning Scientist

Key Skills

Analytical and Problem solving, Data Modeling, Data science techniques, Deep Learning, Effective communication skills, Github, GitLab, HuggingFace, IT, Machine Intelligence, Machine learning techniques, Python Programming, PyTorch

Job Description

The Multi-modal Representation Learning scientists in Machine Intelligence focus on the analysis of various biological data, such as transcriptomic and genomic data, protein sequence and structure, images, and more. We specialize in representation learning, multi-modal diffusion models, and decoder-only generative models to integrate different data modalities and enhance our understanding and predictions. We collaborate closely with our experimental colleagues in early research to improve and accelerate the design of new drug candidates. With a team of scientists located in multiple countries, we strive to deliver the latest advancements in multi-modal representation learning to benefit patient health and global medicine.

The Position

We are seeking an exceptional Multi-modal Representation Learning Scientist to join our team as a Senior Data Scientist and contribute to cutting-edge research in the analysis of biological data and integration of different data modalities. As an expert in representation learning, multi-modal diffusion models, and decoder-only generative models, you will bring your expertise in training deep learning models on diverse biological data and connecting them to other data modalities. In this role, you will work within drug development projects in early research, supporting the development of new drug candidates and the initiation of new drug pipeline projects. This requires close collaboration with other research scientists in multidisciplinary teams and solving ambitious scientific goals. The position also requires close collaboration with our data science and machine learning colleagues across global sites and contributing to the onboarding and development of new computational drug design methods.

Relationships

The Multi-modal Representation Learning Scientist reports to the head of Machine Intelligence. Internal partners include data scientists, specialists, engineers, software developers, technology scouts & partnership developers, system engineers, designers, anthropologists, medical doctors, IT professionals, and others across the US and other countries. External relationships include commercial and academic collaboration partners.

Essential Functions

• Develop state-of-the-art deep learning models for the analysis of various biological data types, such as transcriptomic and genomic data, protein sequence and structure, and images.
• Apply representation learning, multi-modal diffusion models, and decoder-only generative models to integrate different data modalities, perform in silico experiments, and enhance the analysis of biological data.
• Collaborate with multidisciplinary teams to design and execute computational experiments that support drug development/screening projects in early research.
• Stay current with the latest research and advancements in deep learning, representation learning, and multi-modal data integration.
• Communicate findings and results to stakeholders through presentations, reports, and scientific publications.

Physical Requirements

Up to 10% overnight travel required.