Updated: Mar 11
Simon foundation is a non-profit research company it researches computational sciences. It has different departments like neuroscience, quantum physics, astrophysics, mathematics and biology.
Motivation for taking up this field
I enjoy problem-solving a lot and have always been interested in computer science. I plan to go for an MS in CS in the future, so I was looking for a research intern in that area.
I applied for the internship through the company's website; initially, I had to submit my CV and SOP and after the screening, there was an interview. I was asked about my projects in detail during the interview as they wanted to look at my previous research experience along with the knowledge of ML. The group is related to structural and molecular biophysics, the research we did was in computational biophysics, and I was modelling heterogeneity in proteins using cryo-Electron Microscopy (cryo-EM) datasets and Machine Learning. They were looking for someone with a strong background in ML, and here, my second-year research internship experience helped a lot. Apart from my internship, I did a PG level course on Information Retrieval here at IITD as my minor degree requirement, and it had a long project which helped me stand out. Also, as I am in chemical engineering, and since it is associated with this area, it helped a lot.
You need a good level of proficiency in ML with enough projects and coding experience in python. You should have done past research internships and projects. Research internships are quite different from company internships as here you don't have to do competitive coding or crack tests. For research internships, you need to have past experience in related research and good knowledge about the matter; also, you should be able to present the work. Having a publication can really help you, but it is not necessary. I got interested in ML and was learning it mainly through online courses in my second year; after that, I did courses at IIT to get more experience in my third year. The courses here at IITD are quite engaging with tough assignments and exams, so you learn a lot. Also, you need relevant knowledge in Math; while I learnt most of the math as self-study, I suggest you take up courses early on.
My work involved the implementation and development of machine learning methods so as to analyze heterogeneity in the cryo-EM datasets. The cryo-EM data provides information regarding the conformational and compositional variability in the macromolecules and this information can be utilized for 3D reconstruction of the protein structure. Electron microscopy is performed on frozen protein particles in order to generate cryo-EM particle images. As the particles are frozen during electron microscopy, we can only generate 3D structures corresponding to a single conformation. To solve this problem we use generative Machine Learning to predict intermediate protein structures along with the free energy landscape.
This was a work from home research internship, and so while I worked nearly 5 days a week, the work hours were quite flexible. The environment was excellent, and since the group is associated with a lot of researchers worldwide, I worked with quite a few celebrated scientists. My tasks often involved presenting my work, which helped me improve my communication and presentation skills. As I was the only one working in ML in my group, and the intern was online due to Covid, I could only interact with my supervisors.
Suggestions for juniors
When looking for remote research internships, make sure that you are given the required logistical support; especially when you are working with large datasets, you need supercomputer access; while I was given access at Simons, I didn’t get it during my 2nd-year internship and thus, faced a lot of difficulties.
Written by: Tushar Srivastava