Data Science Intern, Medical Imaging

Verily, an Alphabet company, lives at the intersection of technology, data science and healthcare. Our mission is to make the world’s health data useful so that people enjoy longer and healthier lives. 

Our team combines expertise in healthcare, data science and technology to improve the health and well-being of our communities. We are developing the infrastructure and solutions to harness the profusion of health information for good. Our data-driven solutions span three primary areas: research, care and innovation. Programs include Project Baseline - our research initiative to increase participation and evidence generation in clinical research; Onduo - our personalized virtual care platform, which includes connected tools, lifestyle coaching and clinical support; and Debug - our effort to reduce the threat of mosquito-borne diseases by combining machine learning with sterile insect technique. We’re also actively working to combat the spread of COVID-19 through new programs like Healthy at Work. 

Description

Verily’s internship is a paid 13 week program for rising seniors, either undergraduate or graduate students, who are interested in working at the intersection of technology, data science and healthcare. The program is designed for all students, and again this year we encourage students who have been historically underrepresented in this field to explore the program, which is a pathway towards full-time employment within Verily. This includes but is not limited to: Black/African-American, Latinx/Hispanic, Native American, students with disabilities, veterans, and non-binary people.

The PathML team applies machine vision modeling to problems in digital pathology, both through automated disease feature classification, and via our Virtual Stainer Autofluorescence Microscopy research program. The microscopy slides we work with present unique challenges when preprocessing and modeling. High-resolution microscopy images are enormous  (fun fact: we share some code with Google Maps!), we frequently have many spectral channels, and the types of labeling we have can vary wildly from disease to disease, so it isn’t your standard out-of-the-box Convnet problem!

As PathML’s Data Science Intern, you would work on a 13-week project. You will have some degree of ownership and control over the direction of the project, with mentorship available. The project will have both large  engineering and ML experimental components, and presentation of your results will require thoughtful analysis methodology and data visualization choices. Your project will either take the form of a model that we can use towards our mission, or an experiment that helps us better understand our autofluorescence modeling technology.

**Join us for a unique 13 week internship that will take place May 16th to August 12th 2022 OR June 13th to September 9th 2022

Responsibilities

• Designing and pursuing a research problem using our microscopy data. Adapt analysis and experiments responsively as you learn about the problem.
• Contributing code in a professional software development environment. (version control, code reviews).
• Communicate your work through visualizations and presentations.
• Working and communicating effectively with mentors, and peers on the Alphabet Pathology team.

Qualifications

Minimum Qualifications: 

• Experience training Machine Vision models via either classwork or a research position.
• Fluency in Python and NumPy.
• Fundamentals in Linear Algebra, Probability, Statistics, and Machine Learning.
• Completed BS in a relevant field. (Computer Science, Computational Neuroscience, Bioengineering, Statistics, Applied Physics, etc.)
• Currently enrolled in university as a full-time MS or PhD  student with an anticipated graduation date on or before the end of 2023.
• Authorization to work in the United States.

Preferred Qualifications:  

• Experience working with large-scale imaging data, including digital pathology data.
• Experience using distributed computation tools for processing large datasets. (Dataflow, Slurm, Apache Spark, etc.).
• Experience with Tensorflow, Keras, or Pytorch.