Help Patchd build the first ever at home sepsis surveillance system. Sepsis kills 11 million people every single year, but 9 million of those people could have been saved if there had been a way to intervene early. The reason that they weren’t, is that 87% of sepsis begins at home before a patient shows up to the hospital. At Patchd, we are building algorithms to make predicting sepsis at home possible. As a data scientist, you will be tackling time-series based machine learning problems to help build the next generation of algorithms that will provide early prediction for these patients and their doctors.
This role is for a NIH-funded contract with an opportunity for a full time position after if desired. You will be working with electronic medical record (EMR) data to build and validate explainable machine learning models. You will be working directly with the CTO to develop these models for high-risk patients and in the process, will develop an understanding of the pathophysiology of sepsis and the challenges currently in the field. You’ll play a pivotal role in improving our core technology and have end-to-end ownership of your tasks and projects.
In this role you will:
- Train, tune and validate machine learning models using Tensorflow, python and GCP cloud infrastructure
- Clean, organize and manipulate time-series data from hundreds of thousands of patient records
- Use techniques like LIME, SHAP, recent temporal pattern mining (RTP) to develop explainable machine learning models
- Develop tooling for running experiments
We are looking for:
- Experience applying machine learning (in particular deep learning) to health data, time series problems or both
- Experience or a keen interest in healthcare, medicine or biology
- Knowledge of programming and machine learning tools including Python, SciPy/NumPy/Pandas, TensorFlow, Git and cloud computing platforms
- Bonus: Experience with research in a quantitative field like CS, physics, EE, computational biology (Phd, Masters or other projects)
About Patchd
Patchd (patchdmedical.com) is a Y Combinator and VC backed startup, based in San Francisco. Our mission is to help significantly reduce the 11 million deaths that occur every single year from sepsis. We monitor high risk patients who are at home (eg. chemotherapy, end-stage renal disease or organ transplant) for sepsis and when we detect sepsis is likely, we alert the treating team who can intervene. By detecting sepsis outside of hospitals using non-invasive monitoring and AI/ML models, we should be able to change the trajectory of a patient and prevent both death and permanent disability. Robert, the CEO and cofounder of Patchd has personally survived 18 episodes of sepsis and is now trying to protect others from the condition.
Patchd Medical
