Internship

Job description

Job Description

Deep neural networks have revolutionized the way we approach image classification today. The department of Chemical Biology and Therapeutics at NIBR, would like to continue to expand its understanding of deep neural networks for classifying and understanding compound effects from images of cellular experiments. We are focusing our attention on the isolation and quantification of biologically relevant phenomena from features that are captured by the neural networks. We hope to be able to tease out hidden generative factors which can explain the underlying biological phenomena. Visualization of learned features, their occurrences and strengths in the images is critical for our understanding of these features and how they relate to known biology.

The internship will provide an opportunity to experiment with deep neural networks on internally generated, synthetically generated as well as publicly available experimental ground truth data. The focus will be on exploring visualization methods and the disentanglement of generative factors in order to bring us a step closer to high-confidence unbiased phenotype quantification from cellular images. The internship will also provide motivated students with an exposure to pharmaceutical industry and with the opportunity to learn in an exciting, multi-disciplinary and multi-cultural environment with leading scientists and domain experts. The duration of the internship is expected to be between 3 and 6 months.

Desired profile

Minimum requirements

Currently enrolled at a university as a Masters or PhD student or a recent Masters, in Bioinformatics, Biomedical Sciences, Mathematics/Statistics or Engineering. Fluent in English.
Ideal candidate should have a good familiarity with quantitative methods and/or modeling in one-two domains among the following: mathematical, statistical, biological, programming, computational science
The candidate should be proficient in English and have prior programming experience, preferably in Python.
Experience with deep neural network frameworks such as Caffe or Tensorflow is a plus.