First and foremost, it goes without saying that I love working on all things Data Science / Machine Learning.
What fuels me on a daily basis is the ability to work on and solve cool problems in order to help people, learning everything I can all the way to the finish.
My personal interests also extend well beyond what is on my resume and what I do for a living!
I also love studying and reading about anything that interests me in the moment.
I am a musician, DJ, producer, artist and writer.
Music in particular is something that I have cultivated over many years: (Soundcloud / Bandcamp).
nanoDDSCAT+
I developed a tool for the University of Illinois that creates a single workflow for the simulation of light scattering on a custom shape using the open-source code provided by Draine and Flatau's DDSCAT. The tool is open source and everything can be run using a VM in the browser provided by Nanohub.
For this project, I developed a tool designed to allow exploration through simulation of environments characterized by a molecular structure attached to a substrate by an anchor. The code for the tool is built into Visual Molecular Dynamics (VMD) and also makes use of Molecular Dynamics (NAMD), softwares developed by the Theoretical and Computational Biophysics Group at the University of Illinois. The tool is open source and everything can be run using a VM in the browser provided by Nanohub.
This project was the culmination of a merger between the nanoDDSCAT+ and Functionalization Workbench tools. The result is a combined task stream that allows the building of a structure with detail at the atomistic level which can then be put into a simulation for light scattering. Tutorial available at Nanohub.
I wrote a simple tool for creation of game files for rhythm-based video games (something that is typically painstakingly done by hand). The aim of this open source script is to reduce the amount of human effort spent on analyzing songs for relevant sound features. Specifically, relevant features are mapped according to a given time signature and presented in an output format readable by programs such as Stepmania. Time series analysis is powered by the AUBIO library for Python.
Client:
Open Source Community
Year:
2015
Role:
Software Developer
Tera Firma Music Video
As a fun side project, I got myself familiar with Unity scripting and created a music video. Unity 3D provides the perfect basis for creating a 3D world and animating it using time series data. Credit goes to PixelLab for providing starter models for many of the 3D shapes. The animation script extracted sound frequency levels from the song in order to animate different pieces of the environment. Finally, check out the video on Youtube
Cryptocurrency gained a lot of hype in 2016. Given the liveliness of the market and a publicly available exchange API provided by Coinbase/GDAX, I decided to create an automated trader that would buy/sell currencies based on market trends.
Client:
Self
Year:
2017
Role:
Software Developer / Data Scientist
Life360 Data Science
The Data Science products I worked on for Life360 included backend pipelines
to enrich Location data with more insightful features as well as a user-facing product which
could suggest important places or events to users.
Kaggle is a Data Science competition platform that connects Data Scientists of all levels via interesting data sets. Competition hosts make their data public using the platform, along with a specific hypothesis they wish to have answered, and the community responds with an ambitious and knowledgeable atomosphere.
Code used for the competitions is strictly only allowed to be released after it has first been made public to the Kaggle community. After completing two practice competitions, I successfully scored in the top 50% of entrants for the Grupo Bimbo price-demand-prediction competition as well as achieved a medal denoting public recognition for a portion of the code which I released as a kernel to the community during the event.
