Cartographer takes data from multiple IT systems and builds a patient journey including everything from surgical supplies items, to who cared for the patient, to billing data.
Our custom database unifies myriad data sources into one standardized model and our custom query language allows anyone to access relevant data quickly. We use that data to help reduce the data entry burden, provide recommendations for improvements to care, and advance clinical research.
We saved our first client $3M/year, now we are working with eight hospital chains. Started less than two years ago, so we've seen rapid growth. Our product line is applicable to virtually every hospital, so we have lots of growth ahead of us.
Our core platform does some really awesome stuff with NLP to build a rich knowledge graph that tells the story of a patient going through the health system. It solves a number of open data science challenges and could be easily applied outside of healthcare.
We have a really fantastic team of passionate and talented technical folks. For example, our CEO helped discover the Higgs Boson, ran the data platform team of Deepfield, an internet infrastructure company acquired by Nokia, and then went to Stanford's business school to study healthcare.
Our engineering team uses an agile methodology. First, we set a 1-2 month version with larger features based on client interviews and product strategy. Then we pick out Jira tickets for the weekly sprint. As these are developed, our engineers create GitHub pull requests for code reviews. Once the weekly development is complete, we QA and push on Friday. Last, if we will hot fix any bugs if needed.
For new feature development, our product team will work closely with subject experts (for example the nurses and doctors on our team) to develop wireframes and user story tickets. Engineers typically work with our designers to flesh out the specs, then work on the feature. We have a weekly meeting of engineering and product to determine how to develop our major version releases.
Our core data platform, Cartographer, has to talk to dozens of systems, harmonize them, run them through an AI pipeline, and weave the result into a huge knowledge graph about each patient. There are many challenges inherent in doing this, including managing performance, architecting the system to make it as easy as possible to add new sources of data, and doing all of the above on whatever hardware the hospitals are able to provide.
Currently more than 50% of clinician time is going to data entry and paperwork. This is particularly painful when doing clinical research as collecting data from many sources into one common format remains a 6-12 month project at most hospitals. For these research registries alone, our current clients spend more than $6M/year in employee salary to manually do much of this data abstraction.
We have developed human-in-the-loop software called Atlas to speed up data abstraction by autocompleting structured fields and providing suggestions for unstructured fields using NLP. We also make it easy to understand where the data is being pulled from, why we made a suggestion, and a confidence score based on the source of the data. The UI needs to be visually clean, easy to use, and extensible based on the fields required.
We are hoping you can help us build the next version of Atlas. This involves helping our abstractors find data from many sources, improve our 'note viewer' which visualized our NLP suggestions, and develop a 'registry builder' that allows researchers to specify data collection needs for their next research project.
Our underlying assumption is that there is no certain data in the healthcare domain–even the simplest data will have conflicting values in different parts in the system. We embrace this ambiguity by allowing every field in the system to have multiple values, one per source, in addition to a reliability score. This allows us to make an intelligent decision about how to relieve the ambiguity, whether algorithmically or manually by having an abstractor make decisions to create a gold-standard dataset.
This dataset has huge potential to help solve major clinical issues across the hospital, so we are building a series of products. For example, determining how long a patient should recover in each ward is currently tribal knowledge or unknown. Our integrated data allows us to provide clear analytics and suggestions for how to manage this issue, coordinating doctors and allowing patients to know what will happen to them in advance.
We believe that everyone should work at a place where you believe strongly in the impact that you and your colleagues are having on the world. We all care a lot about what we're doing and are therefore highly results focused and collaborative.
Weekly team lunch together, free snacks
Gym subsidy for all employees
Penthouse office in central San Mateo
We encourage and sponsor our employees when they want to learn more about healthcare or software development
Core working hours from 10am to 4pm for office time. We are flexible otherwise as long as our employees end up delivering full-time work.
Full medical and dental.
We offer a $3,000 moving stipend if you are relocated from out of state.
$2000 new laptop subsidy
Interested in this company?
Skip straight to final-round interviews by applying through Triplebyte.