"Before Data Science" Blog

2021 - current

Before you roll your eyes that I have a blog, hear me out. Before Data Science spins off the popular medium publication, Towards Data Science. My goal with Before Data Science is to bring "cutting-edge" data science concepts found in academic papers to the industry. But why a blog? Why don't I just create a github repo with the code? Well, I have 3 responses. First, I want to develop a greater online presence because it brings me out of my comfort zone. Second, publishing weekly posts creates a structure for me to learn that leverages the Feynman technique. Third, if I can start conversations with cool people, I might learn a thing or two.

Skills: Statistics, Machine Learning, Writing


Data Scientist at Tubi

2020‎ -‎ current‎

Tubi is the largest advertising-supported video on demand (AVOD) service in the US. After working there as an intern during my 2019 summer, I decided to return full time as a data scientist. While the role is ever evolving, my main responsibilities are contributing to the A/B testing pipeline, conducting decision science analyses, and building inferential models. Finally, culture is extremely important to me, so I started DS education initiatives which include quarterly hackathons and paper reviews.‎

Skills: SQL/Python, Statistics, Modeling


Volunteer Data Scientist at Learn To Be

2020 - current

Learn To Be (LTB) is a non-profit that provides free online tutoring to underserved youth around the US. I first joined as a tutor, but after speaking with the founder several times, I joined the leadership team as a part-time volunteer. At LTB I have two main functions. The first is to collaborate with leadership to improve the initiatives and structure of the organization. The second is to ensure that LTB is effectively leveraging data, which takes the tangible form of creating KPIs, improving the data tech stack, and developing actionable inferential models.

Skills: SQL, Modeling, Leadership


Data Scientist + Software Engineer at the Schork Group

2020

The Schork Group was mentioned below, but in one sentence the Schork Group published energy reports for investors, mainly focusing on oil and natural gas. Much of the content of these reports is fixed and thereby can be automated. So, to streamline the process, I developed an “Automation Suite” GUI that generates report templates and financial figures with a single click. I’m proud to say that all reports have been automated which gives the analysts significantly more time. The second main feature was a graph generator that allows non-technical analysts to create customized time series plots, greatly expanding the consulting capabilities of the firm.‎

Skills: Python/Kivy, GUI Development, Automation


Marine Species Modeling Senior Thesis

2020

If you can’t tell, I find the oceans fascinating. Moreover, as both a data science and environmental science student, my senior thesis was a unique opportunity to combine my skills and my passion. I was fortunate enough to work with two extremely talented professors: Dr. James Johndrow of the University of Pennsylvania and Dr. Guerra García of the University of Seville. Together we developed explanatory Poisson and random forest models that linked anthropogenic factors, such as pollution or tourism, to marine organism counts. Then looking to forecast these organism counts for reefs in the Caribbean, we prototyped some ARIMA models.

Skills: Python, R, Data Analysis, Time Series Modeling


Data Science + Product Management Intern at Tubi

2020

The Schork Group was mentioned below, but in one sentence the Schork Group published energy reports for investors, mainly focusing on oil and natural gas. Much of the content of these reports is fixed and thereby can be automated. So, to streamline the process, I developed an “Automation Suite” GUI that generates report templates and financial figures with a single click. I’m proud to say that all reports have been automated which gives the analysts significantly more time. The second main feature was a graph generator that allows non-technical analysts to create customized time series plots, greatly expanding the consulting capabilities of the firm.‎

Skills: Python/Kivy, GUI Development, Automation


Marine Species Modeling Senior Thesis

2020

If you can’t tell, I find the oceans fascinating. Moreover, as both a data science and environmental science student, my senior thesis was a unique opportunity to combine my skills and my passion. I was fortunate enough to work with two extremely talented professors: Dr. James Johndrow of the University of Pennsylvania and Dr. Guerra García of the University of Seville. Together we developed explanatory Poisson and random forest models that linked anthropogenic factors, such as pollution or tourism, to marine organism counts. Then looking to forecast these organism counts for reefs in the Caribbean, we prototyped some ARIMA models.

Skills: Python, R, Data Analysis, Time Series Modeling


Data Science + Product Management Intern at Tubi TV

2019

My role at Tubi was centered around improving the product through data-driven insights. On the product management side, I designed A/B tests and supervised the implementation of new iOS features. On the data science side, I created KPI dashboards and developed a Hidden Markov Model to improve user engagement classification. For my final project, I conducted both a graphical and modeling-based analysis on conversion to identify "black hole" features.‎

Skills: SQL/Python, Product Management, A/B Testing


This Website

2019

I developed this website programming in Jade, Sass, and JavaScript, but also used Jekyll with Gulp to streamline the process. My vision was centered around exploration, while still maintaining an intuitive and effective user experience. On every page you will find “easter eggs” that give a little insight into my personality. Hint: you probably missed one on the about page.

Skills: JavaScript, Jade/Sass, UI Design


Binary Classification Independent Study

2019

The binary classification independent study was centered around developing a forecast as to whether a driver will pass an on-road-exam based on driving simulation data. Some of the algorithms studied included logistic regression, random forest, svm, and neural nets. I also covered foundational modeling concepts such as calibration, component analysis, and ROC curves. ‎

Skills: R/Python, Binary Forecasting, Machine Learning


Technical Consultant at the Schork Group

2019

My main role at the Schork Group was to develop a newsletter automation scripts which reduced the editor’s daily workload by four hours. I also helped develop a pitch deck with client profiles to aid the CEO in acquiring partnerships with two large companies. My work with the Schork Group has continued and I am currently working with a Penn professor to develop statistical forecasts for commodity trading.

Skills: R/Python, Automation, Business Development, Machine Learning


Coral Bleaching Project with Reef Check

2018

After partnering with Reef Check, a non-profit that orchestrates the collection of coral reef data, I conducted a statistical analysis to assess reef health after coral bleaching. In short, the data showed that approximately 44% of reef completely recovered after a bleaching event. I am planning to dive deeper into analyzing reef health during my upcoming senior thesis. ‎

Skills: R, Data Cleaning, Regression

"If you've never been diving, please go."‎



CTO + Co-Founder of Fitalyst

2018 - 2017

Fitalyst, the fitness catalyst, was a motivational chatbot that helped users live a healthy lifestyle. My role as CTO was mainly to prototype chatbots using a click-and-point service called chatfuel, then build iOS the application in swift. I also ran several test pilots, conducted market analysis, and wrote the business plan.

Skills: Swift, UI Design, Business Development


iOS App Projects

2014 - 2018

My first iOS application, developed in Objective-C, was a game that quizzed people on topics for my English class. My second iOS app, was a scheduling calendar created for my high school. Throughout this time, I also developed several flappy-bird-inspired games. Finally, after taking an iOS development class at Penn, I created the app behind Fitalyst (a startup I co-founded).‎

Skills: Swfit, Objective-C, UI Design


TA at Wharton Moneyball Academy

2017

At Wharton Moneyball Academy, I was tasked with educating students about baseball analytics techniques in R. I also served as a residential assistant for which I planned events and tended to camper needs.


DraftKings Forecasting Project

2017

I developed a two-part algorithm that forecasted NBA DraftKings points (linear combination of the NBA’s main statistics). The first part was a web-scraping script written in python and selenium that fetched daily box scores. The second part, written in R, was a quantile boosting algorithm that forecasted DK points. The algorithm achieved 83% training accuracy. ‎

Skills: Python/R, Web-scraping, Machine Learning


Investment Intern at Logan Circle Partners

2016

At Logan Circle Partners, I performed statistical analysis of four major oil companies and found optimal buy/sell points after bond ratings changes. I also analyzed four large oil companies’ quarterly reports to develop company profiles. Finally, I presented my findings in both of these areas to my advisor.

Skills: R, Company Analysis


Shark Intern at The Cape Eleuthera Institue

2015

At the Cape Eleuthera Institute, my main role was to assist marine biologists and PhD students in catching and tagging three species of sharks and rays. This involved setting up gear, collecting data from the fish, and organizing the data for later analysis. ‎

"My passion for sharks started here, at CEI."‎