Careers at Tuva Health
In healthcare, we spend 99% of our time cleaning and transforming data and almost zero time generating insights. The Tuva Project is our attempt to fix this -- Tuva or Bust.
Manifesto
Healthcare data has amazing potential in terms of what we can learn from it. For example, it's possible to use healthcare data to answer questions like:
- How common are adverse effects from new drugs like GLP-1s?
- Which doctors are the best at managing complex cardiovascular disease?
- How does cost and clinical quality vary across different providers and geographic locations?
However, analyzing raw healthcare data requires first transforming it into analytics-ready data. This includes:
- Normalization: Standardizing the format and the values in the data so they are consistent across data sources.
- Enrichment: Adding new tables and columns to the data e.g. service categories, encounter types, treatment periods.
- Data Quality: Testing the data to identify problems that will impact analytic results.
This data transformation is necessary because population-scale healthcare data such as claims and medical records are stored in different formats, with tons of data quality problems, and completely lack higher-level concepts that analysts need to answer important questions like those listed above.
Doing this sort of data transformation requires a common set of healthcare data analytics tools. Every data team we've ever worked with has spent enormous time developing the following tools:
- Common Data Model: Most organizations deal with multiple data sources and need a common data model to standardize the schema so the data format is consistent and data analysis code is re-usable across data sources.
- Connectors: Lots of time is wasted transforming raw EHR and claims data sources into the common data model. For example, pretty much every EHR and claims data set has it's own format. Connectors are the code for transforming these raw sources into a common data model.
- Terminology: There are 50+ terminology and reference sets that are needed for healthcare analytics and they are spread all over the internet, maintained by different organizations, and updated on different frequencies. Teams spend enormous time getting these code sets into their data warehouse and maintaining them over time.
- Data Marts: Data analysts need high-level concepts computed on top of the raw data e.g. was this an acute inpatient visit, was it a readmission, does the patient have type 2 diabetes, etc. Without these they can't answer interesting questions from the data. The code for creating these concepts is commonly created in what is called a data mart.
- Data Quality Testing: There are numerous common data quality issues in raw healthcare data and the library of tests needed to catch them is common as well.
As an industry we spend 99% of our time building the tools listed above and using them to transform data. There is hardly any time actually being spent analyzing data.
The goal of the Tuva Project is to fix this. We're open-sourcing tools for transforming raw healthcare data into analytics-ready data. And we're building these tools using modern data technologies that data teams know and love. We want to make it trivial to transform raw healthcare data into analytics-ready data, so data teams can actually start answering interesting questions.
Our vision is that the entire world uses Tuva to transform their data into the exact same analytics-ready format. Not only will we dramatically reduce the amount of time and energy spent transforming data, but every analysis will be completely re-usable. In the future anyone can publish a new analysis and that code will be completely re-usable by anyone else who has their data in the Tuva data model.
As the amount of time and energy we spend transforming data decreases, and the amount of re-usable analyses increase, we think this will result in a dramatic (e.g. 1000x) increase in the amount of high-quality insights being generated from healthcare data. And as data analysts and patients, this is an enormously exciting future that we're very motivated to work on and be part of.
Benefits
We work hard, but we also know you have a family, friends, hobbies, and a life outside the office. That’s why we offer a great benefits package, not just a competitive salary.
Unlimited PTO
Everyone needs a healthy, happy, and productive life. This means you'll need time to rest and recover without restrictive holiday limits.
Flexible, remote working
We want you to work when and where you're most productive, whether you’re an early bird or a night owl.
Fully Covered Health Insurance
Enjoy the fully covered medical coverage for you and your dependents. We've got you and your family covered for all eventualities.
Meaningful equity
You’ll be contributing to the team's success, so you deserve to share in it. Everyone in our team gets a meaningful equity allocation.
Equipment Reimbursement
Get the tech you need to thrive! We offer a reimbursement program for computer equipment, empowering our team to choose the tools that work best for them.
A World Class Team
Work alongside world-class experts leveraging cutting-edge healthcare data technology to drive impactful solutions and shape the future of patient care.
Current Opportunities
Don't worry if you don't see any roles you want to apply for now. Register your interest to allow us to contact you when a suitable role meeting your criteria comes along.