I wanted to participate in this program since a long time. I was really nervous when the organizations were announced. My mind was flooded with multiple questions, including whether or not I’ll be able to do it? Will I find the right project for me? I started shortlisting some organizations based on my skill sets. I had devoted most of time in the year 2018 working on machine learning and deep learning, with a previous experience in JS. I definitely didn’t want to do anything purely web. So after going through many of them, I can down to two organizations - BMI, Emory University and Red Hen Labs. Okay so here’s a checkpoint I’d like you to take note of.
Choosing the right organizations
The common mistake made is going for well-known organisations which have a record of taking a larger number of slots regardless of the interest. The number of slots doesn’t increase your probability of getting selected. Sometimes there’s even more competition due to a large number of applications flowing in! I have made this mistake once when I applied for Outreachy last summer. I applied for just one organisation as it had the most intern positions and it wasn’t even something I really wanted to do. There was quite a competition. Later when I went through the other organisations, I realized there were projects which I’d have loved to work on and had no applications. It is really important to make sure you choose according to your interests!
The project initially caught my interest as it was literally a combo of all the skills I had and all the things I wanted to work on. I’ve been interested in medical data for a long time. This project focuses on browser-based models on such data which intensified my enthusiasm to dig deeper and think of all the ways CaMicroscope could be improved. I’m really happy and grateful to be able to work on this project this summers!
The following are the tasks for me this summer:
- Creating a workflow to allow model developers to allow their model to be run on a selected image or region of interest by the client, to identify cellular features or cancer.
- Add the latest deep learning models researched in the area of digital pathology to caMicroscope to help the pathologists.
- Improve the segmentation app.
- A tutorial to enable users to make models compatible with caMicroscope.