Machine Learning Intern, Multi-Modal Models

Key Skills

Keras software library, Machine learning techniques, PyTorch, RNA-seq Data processing, TensorFlow

Job Description

The Impact You’ll Make

Recursion is revolutionizing the drug discovery process by combining automated assaying of cell biology, diverse data sets, and advanced analytics. We are seeking a research intern who is excited to explore multi-modal data, and push the envelope in understanding and leveraging the information content of different data modalities to make more comprehensive and powerful maps of biology. You will be working alongside a team of Machine Learning Research Scientists and you will be involved in developing new learning algorithms that help us efficiently discover new biological and chemical relationships.

In this role, you will:

• Interrogate different data modalities (including morphological and transcriptomic assays)—characterizing their strengths, weaknesses, caveats and information content
• Implement, train and optimize deep learning models for jointly embedding this multi-modal data
• Quantitatively and qualitatively evaluate multi-modal models and embeddings
• Collaborate with other ML researchers to more fully utilize Recursion’s growing data universe and thereby expand the applicability and utility of our maps of biology.

Recursion is passionate about open science, so you will be encouraged to work towards publishing any interesting results at the top machine learning conferences (NeurIPS, ICML, ICLR, etc.). Join us at Recursion and be part of a team that is reimagining drug discovery through cutting-edge technology and machine learning.

Location:

This position is based at our offices in Montreal, housed at Mila - Quebec AI Institute, the largest AI institute in the world. We prefer in-person collaboration, but will consider remote work for this position. 

The Team You’ll Join 

You will be joining the Machine Learning Research team in Montreal. This team works on developing the next generation of machine learning approaches for drug discovery. You’ll work cross-functionally with machine learning scientists and engineers as well as topic experts who are invested in the creation of data or application of models to accomplish Recursion’s mission.

The Experience You’ll Need

• Currently enrolled in or recent graduate from a Masters or PhD program in Computer Science or related quantitative field 
• Relevant experience as a researcher, including internships, full-time, or at a lab
• Experience contributing research to machine learning communities, including publishing papers in major conferences or journals (e.g. ICML, NeurIPS, ICLR, CVPR, ICCV and associated workshops) 
• Outstanding past projects, publications, and presentations, and a demonstrated ability to communicate results
• Experience using modern technologies to accelerate machine learning (e.g. PyTorch, TensorFlow/Keras, Horovod, etc)
• Experience with cellular images, biological sequence data, or any form of gene expression data (e.g. RNAseq, scRNAseq, L1000, etc…) is a plus
• Experience with multi-modal models or generative models is a plus

How You’ll be Supported

You’ll work closely with our Research Scientists, with regular 1:1s that provide opportunities for support and feedback, and a collaborative work environment. 

The Values That We Hope You Share:

• We Care: We care about our drug candidates, our Recursionauts, their families, each other, our communities, the patients we aim to serve and their loved ones. We also care about our work.
• We Learn: Learning from the diverse perspectives of our fellow Recursionauts, and from failure, is an essential part of how we make progress.
• We Deliver: We are unapologetic that our expectations for delivery are extraordinarily high. There is urgency to our existence: we sprint at maximum engagement, making time and space to recover. 
• Act Boldly with Integrity: No company changes the world or reinvents an industry without being bold. It must be balanced; not by timidity, but by doing the right thing even when no one is looking.
• We are One Recursion: We operate with a 'company first, team second' mentality. Our success comes from working as one interdisciplinary team.

Recursion spends time and energy connecting every aspect of work to these values. They aren’t static, but regularly discussed and questioned because we make decisions rooted in those values in our day-to-day work. You can read more about our values and how we live them every day