Sarah Frost

Hi, I'm Sarah Frost

I am a recent graduate of the Creative Coding Lab in the Computational Media Department at UCSC. I enjoy using software technologies for new applications and thrive building communication and understanding between technical and non-technical individuals.

My Technical Skills are:

  • LANGUAGES - Bash • Python • JavaScript (AJAX, Flask, React) • HTML • CSS • Java • YAML
  • DevOps TOOLS - Git • Github • Docker • Jira • Ansible
  • CLOUD PLATFORMS - Amazon Web Services
  • OPERATING SYSTEMS - Linux • Ubuntu • Mac

What I've Built

Artaffect is an interactive web application that creates a novel engagement with art collections. In contrast with presentations that organize art by genre, medium, or style, Artaffect lets users visually organize a database of visual art in terms of a series of affect parameters that describe the emotional content of the artworks. I am currently working with Dr. Daniel Shapiro to extend this project and develop a novel implementation of style transfer.

Art I Don't Like

I built Art I Don’t Like, a web application using Python and JavaScript. I used machine learning to classify art and recommend new or unfamiliar art to users. I used the WikiArt API to scrape 52,000 images of art to serve as a dataset. I built the front end of the website using HTML, CSS, and JavaScript and deployed the application with a Flask server. I am using an EC2 instance on AWS to host this site. My project was awarded $2400 from IDEA Hub, a UCSC incubator. I presented my work at Museums and the Web 2019 in Boston, Massachusetts.


Deep Painterly Harmonization is an extension of style transfer, and includes a content object which is placed on the style image. The network then harmonizes the style and the content. My group reimplemented the Deep Painterly Harmonization algorithm in Tensorflow, and extend the uses of the algorithm to explore different categories of visual media modification. We wrote a paper in which we discussed the ramifications of style harmonization and style transfer on societal concepts of art, and compared the results of the Tensorflow and Torch algorithms. The project is called ManuMorph because I was delighted with the idea that we would incorporate one of the group members, Manu Mathew Thomas, into famous works of art - ex: the “Manu Lisa” or the “Venus de Manu”.


I built and tested two machine learning models for image recognition to sort waste from dining halls, and achieved 89% accuracy. I presented the models and paper at IEEE Global Humanitarian Technology Conference 2019.


I designed, developed, and deployed GrowAway, an Android app that recommends plant species to grow. Users can find plants based on their location and amount of light in their apartment. I used the OpenWeather API to display local weather information to users, and built a Google Firebase database to store plant data. I developed the application in Java and XML, and released the application on the Google Play Store.

Computational Poetry

What I've Done

DevOps/Cloud Infrastructure Automation Engineer @ NIH

I wrote custom code for customer relationship management software to automate sending questionnaires to biomedical researchers, saving team members 10 hrs/week by reducing repetitive tasks. I also created Jupyter notebooks to access AWS and GCP Identity Access Management functions, team can onboard new clients 25% faster. Finally I compiled technical documentation for cloud service providers and ported documentation to Github Pages.

Database Reliability Engineer, Rakuten

I developed a scalable, dynamic monitoring system for MySQL and Oracle databases for Americas/EU Technology Team. I built this database monitoring pipeline with Prometheus, InfluxDB, and Grafana on Linux OS, and implemented the system for with 2 MySQL masters and 12 slave servers. I deployed source code to Docker containers and set up automation scripts using Ansible playbooks. I wrote Python and Bash scripts for monitoring Linux processes and wrote config files in YAML.

Software QA Engineer I, Mosabi

I was the lead tester for three iterations of the Android app, wrote test plans and tested app on various devices and operating systems. I coordinated with field team to test app in Ghana and Kenya, relayed findings to ensure bug fixes and UI improvements.

Web Development Contractor @ The National Museum of the American Indian

I worked with the Google Cultural Institute to create an NMAI presence on the GCI website. I prepared metadata and uploaded museum objects to the GCI platform. Blurred artworks with copyright issues and annotated museum views with the objects to improve user experience. I created a Google Analytics report that measured increase in web visits from users in South American countries. I wrote a report outlining the value of Snapchat to the National Museum of the American Indian, developed Geofilters to engage museum patrons, produced Snaps consistent with the museum brand and values. I conducted UI/UX analysis of exhibit website for The Great Inka Road: Engineering an Empire and the education module What Would You Do? I wrote quality assurance reports and tested websites on multiple browsers and mobile platforms.

What I've Taught

iD Tech

UCSC ART 101 - Introduction to Programming for the Arts

I was a teaching assistant for ART 101 – Introduction to Programming for the Arts. I worked with the instructor, Wes Modes, to update lesson plans and develop labs to encourage students to think creatively, and problem solve. I prepared materials to ensure that everyone had the necessary information to successfully complete the course.

UCSC CMPM 80A - Accessible Games

In Winter 2020 I was a teaching assistant for CMPM 80A – Accessible Games. The technical experience of students ranged greatly, and many students in the class had no technical experience. As a T.A., I held office hours and tailored information to undergraduate students based on their background – including their understanding of programming concepts, application development, and game development. I also reviewed and graded assignments and managed grades on Canvas.

Tensorflow for Poets Tutorial

How to Make a Gif

How to build a website

The source code for this website is on GitHub.