The Impact You’ll Make

As a Senior Machine Learning Scientist, you will be building foundation models across a range of data types including vision and other Omics that will power a diverse range of use cases in drug discovery. By leveraging your experience in machine learning, software engineering, infrastructure, and data, you will:

• Build and curate large Petabyte scale datasets for training foundation models across multiple modalities
• Develop new algorithms and architectures and train large models on these datasets by distributing across 100s of GPUs
• Explore properties of the resulting representation model on a range of benchmarks
• Develop new benchmarks for interpreting how models are behaving
• Support deploying these models to Recursion’s platform
• Work collaboratively alongside machine learning engineers, researchers, and product managers
  

Location:

This is a hybrid position based on site at our Recursion Canada headquarters in Toronto, Ontario, Canada.

The Team You’ll Join

Recursion has built a unique dataset (over 1 million unique biological perturbations profiled in a high-dimensional biological assay, making up nearly 20 PB of highly-relatable data) and a unique approach to phenotypic drug discovery using our inference-based Recursion Map to predict tens of billions of relationships between disease models and therapeutic candidates. Additionally, we’ve generated large-scale Omics datasets to complement our phenomics data.

You will join a team of machine learning engineers and scientists working on building foundation models that power a range of use cases at Recursion. Our team’s recent preprint, Masked autoencoders are scalable learners of cellular morphology, highlights the success of this approach at capturing biological signals in microscopy images. We’re continuing to push the boundaries of foundation models by leveraging even larger datasets with additional modalities and more compute, leveraging one of the industry's most advanced supercomputers. The capabilities you bring to the team will help us scale our foundation models and develop net-new capabilities that don’t currently exist!

The Experience You’ll Need

• Expert knowledge of deep learning theory and practice including the ability to:
• Curate and build datasets (experience with biological data is a plus)
• Develop net-new algorithms and architectures, particularly leveraging self-supervised learning
• Run large scale experiments evaluating 100s of hyper parameters and scaling experiments to 100s of GPUs
• Interpret and develop benchmarks for assessing model performance
• Supporting productionizing large models, potentially leveraging techniques like quantization and distillationContribute to conference and journal publications
• Tools: PyTorch, NumPy, Pandas, GCP, Hybrid Cloud, Linux, CUDA, Docker, Kubernetes, BigQuery, large scale distributed systems

How You’ll be Supported

• You will join an interdisciplinary team that’s having a major impact internally and externally
• You will contribute to conference publications and have the opportunity to attend major ML conferences
• You will collaborate with other cutting edge teams at Recursion, including Valence Labs at Mila in Montreal

At Recursion, we believe that every employee should be compensated fairly. Based on the skill and level of experience required for this role, the estimated current annual base range for this role is:

• Developing: $215,000 CAD
• Skilled: $226,000 CAD
• Expert: $242,000 CAD

To learn more about our level within levels, click here.

You will also be eligible for bonuses and equity compensation + our comprehensive benefits package for United States based candidates. The range displayed on each job posting reflects target ranges for US new hire salaries and is determined by job, level, and market factors.

During the interview selection process, you will connect with a Talent Acquisition Partner who will be your advocate and ally to ensure you receive the appropriate compensation that meets your needs for your skills, experience, and relevant education/training, while also reviewing our very competitive total rewards package.

#LI-DNI