Fathom is a Series A startup on a mission to understand and structure the world’s medical data. With an engineering team out of organizations like Google, Facebook, Snap, and Twitch, we are starting by using deep learning to structure the data contained within physician notes in order to automate medical coding which is currently a process performed by 125,000 FTEs and costs the US healthcare system almost $10B annually.
Are you passionate about machine learning and looking for an opportunity to make an impact in healthcare? We are seeking an extraordinary Software Engineer, Machine Learning who can not only design machine-based systems, but also think creatively about the human interactions necessary to augment and train those systems. We want to work with teammates excited about taking ground-breaking technologies and techniques to apply to one of the most important and most archaic industries. This is an opportunity to join our team and mission to understand and structure the world’s medical data, starting by making sense of the terabytes of clinician notes contained within the electronic health records of health systems.
Your role and responsibilities will include:
- Developing NLP systems that help us structure and understand biomedical information and patient records
- Using a variety of structured and unstructured data sources
- Imagining and implementing creative data-acquisition and labeling systems, using tools and techniques like crowdsourcing and novel active learning approaches
- Working with the latest NLP approaches (BERT, Transformer)
- Training your models at scale (Horovod, Nvidia v100s)
- Employing and iterating on scalable and novel machine learning pipelines (Airflow on Kubernetes)
- Reading and integrating state of the art techniques into Fathom’s ML infrastructure such as Mixed Precision on Transformer networks
We are looking for a teammate with:
- 2+ years of development experience in a company/production setting
- Experience with deep learning frameworks like TensorFlow or PyTorch
- Industry or academic experience working on a range of ML problems, particularly NLP
- Strong software development skills, with a focus on building sound and scalable ML
- A real passion for finding, analyzing, and incorporating the latest research, technologies and techniques directly into a production environment
- Good intuition for understanding what good research looks like, and where we should focus effort to maximize outcomes
Bonus points if you have:
- Developed and improved core NLP components and not by just 'grabbing things off the shelf'
- Led large-scale crowd-sourcing data labeling and acquisition (Amazon Turk, Crowdflower, etc.)