My Time At Caltech
🛰 Work that was out of this world!
Last updated: December 29th, 2022
Software Development Intern
(Mar '21 to Sep '21)
I joined Caltech after attending a seminar hosted by my high school's astrophysics society, where I met Dr. Kirkpatrick. His research on space really resonated with me, and I was excited to have the opportunity to work with him. When he mentioned that he needed someone with a tech background on his team, I eagerly offered my services. During my internship, I was given three tasks: cleaning up data on brown dwarfs that had been collected over many years, creating a tool for storing and organizing the data, and fixing a program that generated star mosaics but had memory leak issues.
Cleaning the Data, Writing the Paper
My first assignment was to clean up a large and poorly organized dataset on brown dwarfs. The data had been collected over a decade, and there was no consistency in the way it was recorded. Some dates and measurements were in different formats, and there were many missing entries. To deal with this mess, I used the pandas library in python to standardize the data and filter out any rows that had missing or unreliable information. With the cleaned-up dataset, I was able to collaborate with a friend from my high school's astrophysics society, Tarun Kota, on a 📜 Research Paper. In the paper, I helped analyze the data and create graphs to visualize our findings.
Developing the Application
After organizing the data, I set out to create a web application using Django that would allow researchers to easily search and retrieve the data. I hosted the application on an AWS EC2 instance and implemented SSL/HTTPS security and an authentication system to protect the data. I also added a feature that allowed users to export the data into a CSV file for further analysis with external tools.
Star Mosaics
For my final project, I was tasked with fixing a memory leak issue in a software called 🌟🔭Findercharts, which was used to analyze brown dwarfs. After investigating the problem, I determined that the source of the memory leak was a dependency that was no longer being maintained, so I had to come up with an alternative solution. I wrote a Bash wrapper script that would process a set of coordinates one by one, and properly shut down the program to prevent memory leaks. This solution was successful and allowed us to generate more than 3,000 image clusters without any further issues. The project was hosted alongside the webserver which allowed for these image clusters to be queued and generated externally on the cloud. Below you can find a few examples of these mosaics.
