Anomaly is enabling fast, accurate healthcare payments.
Over $300B is lost every year to avoidable denials and payment errors, impacting the affordability of healthcare. We’re partnering with the largest healthcare companies to build novel technology to reduce friction in healthcare payments. Smart Response uses AI to predict denials and payment amounts so providers and patients can avoid expensive denials before they occur. This solution is powered by our claim prediction engine, which analyzes billions of transactions to predict claim payments and denial reasons with over 98% precision. Instant Pay (coming soon) will further streamline payments by enabling providers to immediately get paid at claim submission. Founded in 2020, we recently announced $17M in Series A and seed funding. Our fast-growing team brings together top engineers and healthcare experts committed to making healthcare more efficient and affordable.
Our engineering blog includes more info on the problem we're solving and large datasets we work with.
Applied ML Engineer Role
As an Applied Machine Learning Engineer, you will play a key role in building the prediction engine that powers Anomaly's products to make healthcare payments more efficient. This includes executing cutting-edge ML techniques; training and testing models on hundreds of millions of data-points; and helping build-out a state-of-the-art human-in-the-loop ML infrastructure.
In this role, you will closely collaborate with other data scientists, members of the platform engineering term, and payments domain experts as we build products to streamline healthcare payments and reduce cost and complexity.
You will report directly to the Head of Data Science. The role is open to fully remote candidates who are US-based, with an expectation of occasional travel to our NYC headquarters.
What you'll do:
Write production-level machine learning models
- Executing cutting-edge ML techniques to predict insurance payment amounts, denials, and overpayments with high precision, recall, and explainability.
Making infrastructural choices in order to ensure that the trained models scale well across hundreds of millions of health insurance claims.
Collaborate with the team to build internal tools to support our company mission
- Working with domain experts in the health insurance field to incorporate their knowledge into trained models.
- Building interactive internal tools to obtain expert feedback for the purposes of model validation.
Have strong impact on innovative engineering projects
- Running unsupervised anomaly detection on millions of heterogeneous health-insurance features using PySpark.
- Shining a bright light on black-box models to make all anomalous signals easily interpretable.
- Interacting with domain experts and encoding their experience into simple heuristics. Pushing these heuristics to production for the purposes of weak supervision.
- Training large-scale, weakly-supervised ML models using tools like Snorkel. Making sure these models are interpretable.
- Tweaking models using expert feedback. Building dashboards to solicit that feedback using human-in-the-loop training techniques.
- Experimenting with transformer model applications using the Pytorch / Tensorflow libraries.
What you'll need:
- Standard knowledge of Python data-science stack
- 5+ years in software engineering experience
- Hacker mindset
- Experience building out production-ready machine learning solutions
Nice To Have
- Exposure to PySpark, PyTorch, Snorkel
- Experience working in a fast-paced, innovative environment
Apply here, or reach out to firstname.lastname@example.org with questions.
Life at Anomaly
Headquartered in NYC with a strong remote team, Anomaly brings together a diverse group deeply committed to our mission to bring streamline healthcare. Data scientists, engineers, clinicians, and more work together to realize a future where healthcare payments flow with precision. We live by our values, which help us accomplish more and create a workplace we love.
Move with urgency: Our customers needed our product yesterday. We move fast—but not too fast—to bring our solution to market
Be willing to sit on the floor: No one is too senior to jump in and get the job done
Think 10x: We celebrate incremental improvements, but we always look for the step-function win
Seek feedback at 30%: We ask for feedback uncomfortably early, and offer it proactively (and respectfully) to others
Default to open: We internally share everything we can, creating a culture of trust and empowerment
Be kind: We take our work seriously, but we’re never too busy to be kind