Freelance
I do freelance work through a number of online platforms as well as working directly with the client. Links to my profiles are included below; feel free to browse them and contact me if you have any work related to data science, machine learning, or general software design. I generally prefer to work outside of either platform, and if you feel the same you can contact me directly over email.
Details about some of the work I've done for clients as a freelancer (the one's I thought were cool enough to share) is included below. Details about most of these are kept vague due to NDAs, but you should be able to get the general idea about most of them.
Project Manager - O'Donnell Learn
I've worked with O'Donnell Learn multiple times to lead teams of subcontractors on projects ranging from designing online Computer-Science courses, to migrating traditional courses to online platforms for universities, to working on projects with large companies (like creating "Qwiklabs" and JavaScript widgets for Google's IT Support Coursera Certificate).
I've worked with O'Donnell Learn multiple times to lead teams of subcontractors on projects ranging from designing online Computer-Science courses, to migrating traditional courses to online platforms for universities, to working on projects with large companies (like creating "Qwiklabs" and JavaScript widgets for Google's IT Support Coursera Certificate).
Data Analysis for Supply Chain Improvement
I'm working with an international shipping company specializing in perishable produce to analyze data they've acquired in order to improve their shipping process. We're analyzing their data and doing some machine learning to identify ways to limit the amount of shelf life that the produce loses during transit, providing static analysis and data visualization to inform client decisions, as well as real-time predictions of end-of-flight quality.
I'm working with an international shipping company specializing in perishable produce to analyze data they've acquired in order to improve their shipping process. We're analyzing their data and doing some machine learning to identify ways to limit the amount of shelf life that the produce loses during transit, providing static analysis and data visualization to inform client decisions, as well as real-time predictions of end-of-flight quality.
Product Recommendation TwitterBot
I created a TwitterBot (using Python and Tweepy) to crawl through Twitter and identify users that are likely to be interested in a certain type of product based on their tweet history. When users are matched to a product category with a high-enough confidence, a product within that category is retrieved from Amazon and recommended to the user.
I created a TwitterBot (using Python and Tweepy) to crawl through Twitter and identify users that are likely to be interested in a certain type of product based on their tweet history. When users are matched to a product category with a high-enough confidence, a product within that category is retrieved from Amazon and recommended to the user.
Prototype "Fact-Checking" API
As a prototype for a client that was exploring the possibility of creating a startup, I created an API (using AWS API Gateway, Lambda, DynamoDB, and S3) that scores documents or articles based on how "trustworthy" they are. It does this by having a bank of "trusted" articles that are separated by topic, and looking for similarities between the trusted documents and new ones at the sentence level (using Python and NLTK). Based on user-specified thresholds for what is considered a close match, it pairs sentences in the new document to trusted sentences and returns an overall document score (as well as the matching sentence pairs).
As an example, verifying a small paragraph about Shakespeare against a few "trusted" articles about him results in these matched sentences:
Original Sentence: "Twins Hamnet and Judith followed in February 1585."
Matched Sentence: "On February 2, 1585, twins were baptized, Hamnet and Judith."
As a prototype for a client that was exploring the possibility of creating a startup, I created an API (using AWS API Gateway, Lambda, DynamoDB, and S3) that scores documents or articles based on how "trustworthy" they are. It does this by having a bank of "trusted" articles that are separated by topic, and looking for similarities between the trusted documents and new ones at the sentence level (using Python and NLTK). Based on user-specified thresholds for what is considered a close match, it pairs sentences in the new document to trusted sentences and returns an overall document score (as well as the matching sentence pairs).
As an example, verifying a small paragraph about Shakespeare against a few "trusted" articles about him results in these matched sentences:
Original Sentence: "Twins Hamnet and Judith followed in February 1585."
Matched Sentence: "On February 2, 1585, twins were baptized, Hamnet and Judith."
Data Visualization Tools
Many clients require custom data visualization tools to accompany their projects, either as part of projects I've worked on or to supplement their own. All of these need to generate the plots, charts, etc. on the fly based on the input they're given, and be as visually-appealing and easy-to-understand as possible. Below are a (very) few samples of selected visualizations from some of my previous projects for clients (all data has been obfuscated):
Many clients require custom data visualization tools to accompany their projects, either as part of projects I've worked on or to supplement their own. All of these need to generate the plots, charts, etc. on the fly based on the input they're given, and be as visually-appealing and easy-to-understand as possible. Below are a (very) few samples of selected visualizations from some of my previous projects for clients (all data has been obfuscated):