Staff AI Researcher

The Staff AI Researcher is responsible for developing advanced artificial intelligence solutions
that improve health outcomes for millions of patients by empowering primary care physicians
with technology that keeps patients healthy and prevents unnecessary hospitalizations. The Staff
AI Researcher collaborates with engineering and analytics teams to bring AI technologies into
existing products and workflows. Additionally, the role involves training, fine-tuning, and using
AI models harnessing knowledge from extensive data sets of medical records, diagnoses, claims,
and prescriptions collected from millions of patients across the country. Primary duties include
building working prototypes using off-the-shelf and novel AI techniques to deliver higher levels
of optimization for the company; working with large, complex data sets and solving difficult,
non-routine analytical problems to harvest data; redesigning existing pipelines and systems to
meet growing data and query needs; implementing techniques for fine-tuning and adapting pre-
trained generative models to specific healthcare domains or tasks; developing evaluation metrics
and benchmarks to assess the quality and performance of AI/ML models; designing and
implementing feature engineering pipelines, including data processing, feature extraction, and
transformation to optimize model performance; setting and upholding standards for engineering
processes, including style and code checking, test harnesses, and release packaging; and
delivering working proof-of-concept solutions that balance speed, scalability, and time-to-market
considerations.

Must have a Master’s degree or foreign degree equivalent in Computer Science or a related
quantitative field and six (6) years of machine learning and statistical analysis experience. The
position also requires demonstrated knowledge and experience with the following: three (3)
years of deep learning and large language model experience; three (3) years of Python
experience; three (3) years of proficiency in selecting the right tools given a data optimization
problem; addressing challenges from incomplete, unrepresentative, and mislabeled data; and
large-scale distributed systems at scale and statistical software (e.g., Spark).