Nachiket Subbaraman

As a software development engineer at Amazon Web Services, I develop and optimize machine learning solutions for personalization and recommendation systems. I work with PySpark, Glue, Java, and Python to create scalable and reliable cloud-based applications that serve thousands of customers worldwide.

I graduated from the University of Illinois at Urbana-Champaign with a bachelor's degree in Mathematics and Computer Science, where I gained a solid foundation in data structures, algorithms, and software engineering. I also completed several internships at Qolsys and Loadstar Sensors, where I automated testing, architected scripts, and improved metrics. I am passionate about learning new technologies and solving challenging problems that have a positive impact.

• Apache Spark, Large Scale Infrastructure, Machine Learning, Operations, Code and System Health, Testing, Debugging, Documentation, Concurrency, Multi-Threading, Data Processing, Eager to Learn and Curious, Ownership

● Research Assistant with Professor Richard Sowers, Math Department, UIUC                                                           

○ Constructed algorithm to store large amounts of hierarchical web scraped data in a tree data structure

○ Implemented KMeans clustering algorithm to plot wildlife data and compute density metrics

○ Parallelized time-consuming function which computes entropy and density, dividing runtime by number of CPU cores

○ Documented important milestones from GitLab repo commit messages and email threads

● Research Assistant at Forward Data Lab, UIUC (https://github.com/nachsub/Keywords-Forward)                                                    

○ Created a searchable index of a corpus of computer science research papers via Whoosh, a search engine Python library

○ Populated another searchable index from which clients can enter search phrases on a Flask web application and retrieve a dictionary ranking of related keywords based on NLTK, a natural language processing Python library

● Illinois Geometry Lab Research Project (https://tinyurl.com/yapdbyrj)                                                                                                              

○ Collected data sent for storage in an AWS S3 bucket using AWS Firehose as a delivery mechanism.

○ Transferred the S3 data to AWS Redshift for analytical processing using the SQL copy command. 

○ Displayed water depth vs. time data on a web browser using a local Node.js server connection to the Redshift database.

● PCA Classifier (https://github.com/nachsub/cs361sp20/blob/master/cs361_final_project_part2.ipynb)  

○ Implemented PCA plots on sample data of eigenvalues in sorted order and top two principal components  

● MNIST Classifier (https://github.com/nachsub/cs361sp20/blob/master/cs361_final_project_part2.ipynb)  

○ Implemented MNIST classifier via PyTorch trained neural network 

● Optimization (https://github.com/nachsub/cs361sp20/blob/master/cs361%20project%20code.ipynb) 

○ Implemented stochastic optimization algorithms for gradient descent ADAM, SGD, and ADAGRAD in Python using Numpy

○ Plotted training loss of each gradient descent algorithm using Matplotlib error bars

● Shell (https://github.com/nachsub/cs241_sp2020/tree/master/nks5-master/shell)    

○ Created a Linux shell in C that executes built-in and external commands using fork/exec/wait and signal handling

● Brick Breaker (https://github.com/nachsub/cs126-final-project)    

○ Developed a C++ OpenFrameworks game that updates a Firebase database via HTTP JSON and GET requests.

● Hack Illinois (https://github.com/openreferral/hsds-transformer)    

○ Added functionality to open source repository for users to input a file path for the output directory in Ruby

● Diner Finder (https://github.com/nachsub/DinerFinder)    

○ Find nearby restaurants using the Google Maps API via Android Studio application development

● Tick Task (https://github.com/ERiverIllini/TickTask) 

○ Prioritizes assignments by estimated time to complete and due date using MongoDB, Express, React, and Node.js

○ Updates events from users’ calendars to their website via Node.js script 

● Naïve Bayes (https://github.com/uiuc-sp19-cs126/naivebayes-nachsub)   

○ Classifies a text file of images of digits, each digit of size 28x28, as digits 0-9 through windows console output using the Naïve Bayes machine learning classifier in C++