Arnav Shirodkar.
I'm a Software Developer
Percussionist, Composer, and Music Tech enthusiast
Hi, my name is
About Me
I am a software developer from Singapore with a passion for exploring the intersection of music and technology. I also have a strong background in percussion performance and electronic music.
I am currently working on my Senior Project, a symbolic music transformer model for real-time improvisation alongside a live performer. Built with MaxMSP, Node.js and Typescript for the performer interface, Python and Tensorflow/ Pytorch for the transformer model.
Here are a few technologies I've been working with recently:- PyTorch
- Typescript
- Shiny (R)
- Max/MSP
- Tensorflow
- TidalCycles
Experience
I’m currently working on a research project with Prof. Valerie Barr as part of an NSF-funded Alliance for Identity Inclusive Computing Education (AIICE) project to analyze CS degree completions for intersectional identities across all Title-IV institutions in the U.S.
- Designed and implemented a Python data scraper as well as an interactive web app with R to show trends in degree completions with ggplot2, handling upwards of 10 million data points across 20 categories using dplyr and meeting web accessibility standards.
- Built an automated workflow with Bash and Python to seamlessly test and integrate data scraping, cleaning, processing and Shiny app deployment for data released in subsequent years
- Tutored Mathematics & Computer Science weekly for incarcerated students enrolled in the Bard Prison Initiative (BPI) at Eastern and Fishkill Correctional Facilities, creating custom assignments to help develop proficiency in Data Structures and Algorithms within the prison environment.
- Served as an in-class tutor for over 50 students at Bard College in fundamental CS classes (Data Structures etc.), leading pre-exam review sessions as well as 1-on-1 tutoring for struggling students.
- Also served as a CS Instructor for CAMP 2023, an educational summer program for middle school students focused on Mathematics and Computer science in a creative intellectual environment
- Worked with research and quant team to lead the migration of outdated data quality tests to Morningstar’s new data quality framework, improving test coverage for updated data schemas to 100% and improving test execution times by 20%
- Contributed changes to Great Expectations (v2), an open source data quality framework, adding support for pairwise data column testing for SQLAlchemy backends.
- Deployed a new AWS pipeline in Morningstar’s Data-Lake Eventing framework (Athena, SES, SQS) using Docker, Jenkins, Typescript and Splunk-On-Call to detect and report critical test failures to data producers, improving failure response times by 40% for ~ 15 teams
- Led early development of a new in-app chromatic tuner by analyzing user statistics and community engagement forums for company’s iOS app before proposing the new feature, enhancing product usability for 30% of the customer base
- Created a rich tuner design with multiple controllers, custom components and reactivity across various device screen sizes using Swift and iOS storyboards for a cohesive user experience
- Implemented real-time pitch detection using signal autocorrelation and iOS Core Audio, maintaining original dependencies to ensure backwards compatibility with iOS 7.0 and above.
- Built a full-stack dashboard with React, Flask and SQLAlchemy for Musiio’s assistance in Vans’s “Musicians Wanted” competition, tracking 4000 song submissions from hundreds of artists across 13 countries
- Improved the efficiency of internal Python scripting with data aggregation, multi-threading and caching: 33% faster YouTube audio tagging, 75% fewer requests to Spotify API
- Delivered 30+ high-value client POCs worth several thousand dollars in revenue, including Universal Music, CDBaby, Hipgnosis, Soundcloud
Education
2019 - 2024
B.A. Computer Science & B.M. Percussion Performance
Bard College & Conservatory
GPA:
3.95 / 4.00
Extracurricular Activities:
- One of 3 Founders of the Bard Computer Science Club
- Captain for Bard’s Ultimate Frisbee Team
High School
National Junior College (Singapore)
A-Level Subjects:
- H2 Mathematics
- H2 Physics
- H2 Chemistry
- H2 Music
A-Level Score: 90/90
Projects
Node.js
Typescript
PyTorch
Tensorflow/ Tensorflow.js
Max/MSP
Real-Time Music Transformer
A symbolic music transformer model for real-time improvisation alongside a live performer.
Python
Matplotlib
TSP Path Visualizer
Created an animated path visualizer for the TSP problem with greedy and genetic algorithm solutions
Python
Tensorflow
Librosa
Numpy
Music Meter Classifier
Designed an end-to-end musical meter classifier using CNNs that inferred rhythmic groupings from mel-spectrogram data in Tensorflow, achieving 73% accuracy
Vue
Javascript
Cordova
Scandit
Quasar
Green Bean
Prototyped a hybrid-native mobile application that provides users with sustainability information for grocery items leveraging barcode scanning libraries built by Scandit, encouraging sustainable grocery purchases.
Tensorflow.js
PoseNet
Flutter
BackBend
Prototyped a mobile application for a hackathon that uses mobile device cameras to monitor and pings users with bad posture while they work.
Achievements
Distinguished Scientist Scholarship
Receipient of the Distinguished Scientist Scholarship from Bard College
John W. Boylan Scholar in Medicine and Science
Receipient of the John W. Boylan Scholar in Medicine and Science from Bard College
Morningstar Tech Case Study Winner
In a team of 5, developed an interactive trading simulation that tracks user behavior to measure indicators of investor overconfidence
Techstars Startup Weekend Runner Up
Pitched an idea for an AI based cybersecurity startup to a panel of judges and investors, winning 2nd place out of 9 teams
Get in Touch
Feel free to reach out to me for any opportunities or collaborations!