Single Cell Representation Internship

Key Skills

Algorithms, Analytical and Problem solving, Bioinformatics, Deep Learning, Linear Algebra, Machine learning techniques, NeurIPS, NoSQL, Python Programming, SQL

Job Description

Valence Labs is the AI research division within Recursion, focused on the industrialization of scientific discovery to radically improve lives. We blend the innovative culture of academia with the resources and structure of industry, driving the development of autonomous systems that are changing the way we discover and develop treatments for complex diseases. Our work is guided by a commitment to open science, with team members presenting at top-tier conferences and contributing to some of the largest open-source communities. Our offices are based in London and Montreal, with strong ties to Mila, the world’s largest deep learning research institute.

About the Role

We are seeking an intern to work with our Single Cell Representation efforts to advance critical methods for single-cell data filtering, a foundational step in accurate data analysis for gene knockout studies. This role involves developing and applying computational techniques to remove proximity bias and identify true single and double knockouts in complex genetic perturbation experiments for genome-wide applications. This role offers the opportunity to make a meaningful impact on our mission of improving the resolution of microscopy data analysis with single-cell precision, a critical component in advancing biological discovery.

Responsibilities

• Develop and validate filtering methods to exclude proximity-biased and non-transfected cells in single-cell datasets.
• Apply single-cell filtering methods to pairwise perturbation data, focusing on detecting double knockouts accurately.
• Assess and quantify the effect of filtering on data quality, particularly on image phenomap generation.
• Scale filtering methodologies from single chromosome arms to whole-genome applications.
• Document workflows and findings for internal knowledge-sharing and potential publication.
• Integrate developed methods into the data pipeline to support ongoing research and meet key project milestones.

Outcome Goals

• Deliver a validated filtering method ensuring high purity of double-knockout cells.
• Compile and present progress reports or publish findings on single-cell filtering techniques and their biological significance.

Ideal Candidate Profile

• Current Master’s, PhD, or post-doctoral researcher with a focus in machine learning, bioinformatics, or computational biology.
• Strong programming and software development skills, especially in Python.
• Experience with machine learning and statistical techniques for data analysis.
• Engineering experience in building and deploying high-performance implementations of deep learning algorithms.
• Demonstrated capability to understand and summarize scientific content and implement deep learning models based on descriptions from publications.
• Proven ability to communicate scientific concepts effectively in written and verbal formats.
• Strong knowledge of linear algebra, calculus, and statistics.
• Passion for applying ML research to real-world problems.

Nice to Have

• Authorship of a publication in peer-reviewed ML conferences (e.g., NeurIPS, ICML, ICLR, or similar) or biology-focused journals.
• Experience with open-source contributions in ML/biology libraries.
• Familiarity with cell biology and genetic perturbation experiments.
• Scientific knowledge of biology, chemistry, or physics along with previous experience working in a scientific environment across disciplines.

Valence Labs is dedicated to fostering a collaborative, inclusive environment that prioritizes innovation, open dialogue, and mutual respect. Join us to contribute to groundbreaking work that is shaping the future of scientific discovery.