Muuse is a live music streaming platform designed as a social network that allows anyone to stream high-quality audio & video to their audience. At Muuse I was the sole backend engineer and handled everything from API design to CI/CD and our K8s architecture. Peak usage had 8 concurrent video streams with 1k+ viewers. ⇨ Implemented Flask API utilizing GraphQL & SQLAlchemy.⇨ Designed and managed PostgreSQL database.⇨ Integrated & optimized NGINX-RTMP streaming servers.⇨ Configured staging & production K8s environments and established critical autoscaling criteria for streaming and API servers.⇨ Incorporated CI/CD pipelines for deployments to our K8s staging and production environments (Client, API, Streaming Servers).⇨ Efficiently utilized Digital Ocean hosting to manage DB, K8s, and storage.

At Tatari, I build data pipelines and architectures. I explore ways to better understand and measure user interaction and client performance. I develop and deploy algorithms for internal and external use.⇨ Traffic Anomaly Detection: Developed and deployed a traffic anomaly detection algorithm. This model uses a sparsely labeled data set and supervised learning (XGBoost) to identify when a client is experiencing unusual traffic. This algorithm outputs crucial daily results, enabling Tatari to ensure accurate measurement.⇨ Impression Estimation: Researched and implemented a model that imputes missing impression values from our data sources. This model utilizes hypothesis testing to fall-back to subsequent data source measurements. Designed and built a data pipeline that incorporated the estimation model.⇨ Genre Analysis: Developed an algorithm that allowed clients to understand their performance by specific program genres. This involved building an ETL pipeline for the data source containing genre information and testing various mathematical models to accurately weight each genre.

Completed intensive Data Science Immersive course that solidified skills in Programming, Statistics, and Machine Learning. Completed Several Case Studies and Final Capstone Project.⇨ Smunch User Segmentation and Meal Preference Analysis: Created a comprehensive user analysis with two main functions. One: cluster the users based on their ingredient and nutritional preferences and identify the size and key ingredients for each unique cluster. Two: predict the “successfulness” of a meal based on each customer's learned ingredient preferences. This prediction is wrapped up in a Web Application that Smunch can use to test meal ideas, explore customer preferences, and understand their quickly expanding customer base.⇨ Fraud Detection: Analyzed event postings to give a soft classification of the event being fraudulent. This study utilized a Random Forrest model and NLP techniques to make a prediction. The final product was wrapped into a web app built with Flask and hosted on AWS.⇨ Movie Recommender: Built a recommendation engine that takes user and movie information and ranks top predicted movies for a particular user. The final model utilized ALS Matrix factorization for movie rating predictions.⇨ Churn Prediction: Developed Gradient Boosted model to predict users that are at risk of churning. Validated model with ROC plots and then used the model to create Profit Curves for hypothetical business solutions.

⇨ Implemented a linear regression model to predict the total number of expected meal orders per day.⇨ The model was deployed on the backend website for the operations team to use with interactive parameter selection.⇨ Completed projects such as retention rates, company cohort breakdowns, and informed CEO of specific trends.

Allthenticate is a cross-platform ubiquitous mobile phone security system for physical and cyber security.⇨ Designed MAC OS desktop application with Qt.⇨ Began iOS Application development using Swift.⇨ Helped develop a business model, financial projections and presentation pitch.⇨ For more information on product visit Allthenticate.net

Member of the Photon Control and Data Systems group for two summers completing several engineering projects.⇨ Designed and installed a vacuum valve fail-safe switch with pneumatic monitoring capabilities. It was implemented in experimental hutch and the design was expected to be installed in several other hutches.⇨ Wrote serial driver protocol file for sample delivery system in experimental hutch.