Hello! I'm Abhinav Bandari, a senior studying computer science at the University of Washington's Paul Allen School of Computer Science and Engineering.
I've worked on computer science research, contributed to open source projects, worked on independent projects, and participated in hackathons. See more below!
Developed and shipped feature to efficiently store and query IPv6 addresses in radix tree map.
Built anomaly and outlier detection algorithms for round-trip time of packets sent over the network, using Elasticsearch and Kibana.
Developed automated operations dashboard for Platform Engineering team using AWS CDK.
Integrated metrics from Splunk Infrastructure Monitoring and AWS CloudWatch, Support, and Health services.
Developed Tableau Web Data Connectors using AWS SDK, REST APIs to automatically pull metrics data.
Developed non-contact acoustic sensing algorithm to detect and identify a user's exercise through their smartphone with 83.0% accuracy.
Implemented ultrasound sensing through phone's speaker and microphone, used Fourier analysis to develop audio spectrograms for each exericse, and trained deep neural networks to analyze the spectrogram images and identify the performed exercise.
Worked with graduate and undergraduate student researchers; advised by Dr. Shwetak Patel.
Taught weekly after-school classes to elementary school students in Scratch and Python programming languages.
Worked with another instructor to teach provided curriculum and adapt lessons as necessary.
Developed Android application and React web app to crowdsource anonymous user location data.
App displays real-time heatmap visualization of crowd density, overlayed on a Google Maps interface.
In the COVID-19 era, users will benefit from having real-time information on which locations are safe to visit.
Built during UW Hack'20 hackathon in a team of four members; finished top 3 out of 81 projects.
View source code here.
Contributed to open-source automatic machine learning (autoML) library that automates machine learning preprocessing, model building, training, and evaluation.
Personal contributions: Developed backend for image classification (convolutional neural networks) and artificial image generation (generative adversarial networks) queries.
Library has over 17,000 downloads. View source code here
Developed deep learning classification algorithm to take a chest X-ray image as input and output a diagnosis for four heart and lung conditions (pneumonia, pneumothorax, cardiomegaly, pulmonary edema)
Developed image classification algorithm using dataset of chest X-rays released by the National Institutes of Health.
The algorithm's performance was comparable to performance of state-of-the-art algorithms on this dataset.