Get in touch at hi@rowan.earth
MiraDx (2021 - 2023)
MiraDx is a molecular diagnostics lab with specialty cancer tests. I was hired as the first in-house engineer, built and managed an in-house engineering team of 3, and worked closely with the founder/CEO to identify and decide direction of engineering needs.
I owned the Typescript+Nextjs platform and worked with UCLA scientists to productionize the statistics model in Python. I brought in security researchers to harden the codebase's security, redesigned the system for new markets, and productionized research classification models.
Stanford Department of Medicine, HealthRex (2020 - 2021)
Our goal was to build a better clinical decision support system by training models on as much real clinical data as possible. Health data is notoriously difficult to obtain, but the idea was that nascent techniques like multi-party computing (MPC) could enable training a model without getting access to the data being used to train it.
As part of a larger experimental MPC platform, I built a Python package for transforming medical data into a format usable by models, deployed it along with the multi-party computing ML framework on a few Google Cloud virtual machines, kicked the tires, and wrote up some documentation for the researchers.
Firewall Cafe (2020 - 2021)
Firewall Cafe is a project that began as a gallery installation that allowed people to simultaneously do an image search on Baidu and Google as a way of exploring censorship of the results. All searches and images that are discovered are saved into a dataset. But having a dataset that was collected in a systematic way, tracking terms of interest over time, might prove to be interesting for researchers and journalists looking into censorship. So I build Great Firewall Watch, which runs at a regular interval on a specified set of terms. I prototyped the process in a series of notebooks, and built a portal for researchers to browse its results.
FiveThirtyEight article, freelance data journalism (2020)
I was a contributing researcher to an article that appeared in FiveThirtyEight on COVID's effects on naturalized citizen voting, and published the data I collected as a GitHub repository. We looked at the relationship between naturalization offices closures and effects on voter population, with an eye towards the upcoming 2020 elections.
Thespie (2019 - 2020)
I was the back end lead at Thespie, a theatre curation service. I built a recommendation system on an AWS stack, a similarity ranking system, and was responsible for the API.
Recurse Center, Winter 1 2018 batch
I did a 12-week batch at the Recurse Center in downtown Brooklyn. My focuses were deep learning (especially the fast.ai courses), natural language processing, geospatial data, and computational linguistics. My final project was to collaborate with a linguistics professor to build software that transformed Wikipedia articles into maps.
Rheos Medicines (2018)
Created a dashboard and data pipeline for visualizing data from whole-metabolome experiments, inspired by ReconMap.
The metabolome was represented as a massive, static graph layout with ~20k nodes and ~20k edges. A Python pipeline colored nodes based on data and rendered tiles for a front end using a custom implementation of Google Maps, which ended up being by far the fastest approach in terms of both reactivity and developer time.
NuMedii (2017 - 2018)
Scientists at NuMedii needed a way to explore relationships between drugs, genes, and diseases at a high level, informed by NuMedii's drug enrichment analysis software. Testing drugs for treating diseases is expensive and time-consuming, and the potential exists for existing drugs that have already undergone some testing to be repurposed for treating another disease.
The dashboard was implemented as an Express + D3 front end, making queries to get scientific data from a back end database. I worked on the front end, optimizing load times and add/remove calls to the graph network, building new node types based on other enrichment analysis data, and improving the layout engine.
2016 - 2021 Consulting and contracts specializing in science-centered web apps, data analysis / visualization, and product development.
Microsoft: Updated Microsoft School Data Sync OneRoster API demo to conform to the OneRoster 1.1 spec from IMS Global, an education standards consortium. C#, Azure, .NET, Micosoft SDS API.
GoChip: Developed core features of desktop and phone apps brought to market. Built core features and made UI improvements. Node, Electron, React and React Native.
edX education platform: Contributed to user accessibility in Django.
personal project (2016)
This map project is a love letter both to D3's wonderful mapping tools and the biodiversity in the Pacific Northwest. Read the longer making-of here, but in short, I built a dashboard that uses tree range maps to indicate the ecosystems where various evergreen species grow in Washington state.
Center for Game Science, University of Washington (2012 - 2015)
I was part of a small team of research engineers at the Center for Game Science that built Nanocrafter, a game which challenged players to build machines out of RNA to solve puzzles which was funded by DARPA and the National Science Foundation. I also spearheaded the design of its peer scoring, which later won the game an award at the Serious Games Showcase and Challenge. Sadly, the game was shut down a few years ago, but is hosted by the Internet Archive.
- Bachelor's of Arts in Computer Science from St. Mary's College of Maryland, 2012
- Summer Intern at Sandia National Labs, Complex Adaptive Systems of Systems Engineering initiative, 2012
- Summer Intern with the University of Michigan + Google Summer of Code, Center for the Study of Complex Systems, 2011