Skip to content

Avyntra

Kent Reese edited this page Jun 7, 2019 · 4 revisions

Background

Elevator Pitch

Location: India

Avyantra addresses the early diagnosis of neonatal sepsis by applying a machine learning model that captures data of newborns with neonatal sepsis conditions and their associated treatments.

Initial Call Notes

Current Milestones and Developments

Avyantra is currently developing algorithm. Research for neonatal in India is extremely limited right now. So, they are developing an application for collecting data in addition to the machine learning algorithm. They've identified 70-75 variables, 2 hospitals are currently on board and collecting data already. Writing algorithms now, positioning the entire thing as a storage for medical records in the future. They hope to extend this to other neonatal diseases.

Current Open Source Development

  • They have an all in-house team, data collection, data science part, and have a partner for application development. All tools are all open source, public repository for the code. Two repos, one for application and one for the algorithm.
  • They are looking for guidance on getting contributors to their project.
  • No prior experience in open source, all experience in proprietary.
  • They do have an upstream, post issues in the upstream that they find.

Project Management

  • They use a hybrid of agile and waterfall.
  • Daily SCUMM meetings, daily open floor meetings, weekly management review.
  • Record the minutes, update the task documents regularly, but the documents aren't in the public yet.
  • Internal post-it note board is used for task management.

Testing / Code Health

  • Unit tests and functional tests are both in.
  • 80-90 percent is covered.
  • Right now, testing is just done by whoever can do it.
  • Plan to use a couple health checking and automation tools, testing their usage right now.
  • Peer reviews are already integrated in the organization, also try to bring external experts to do reviews, bringing in a data scientist to review their algorithm code going forward.

Documentation

  • No readme, all project documentation is in their own project folders. A few documents, those required for setup is in the repo.

Additional Notes

Looking for technical expertise in Open Source, strategies for becoming for transparency, what are the requirements for open source, and the code that is being developed. Use cases of other projects doing things successfully.

Their end goal is to be a part of a larger community and get a new experience with open source tools.