Job Description

We are hiring a Senior ML Engineer to advance machine learning on the Empower
Healthcare Platform.

Note on Role Type: This is an **Applied Modeling** position, rather than an
MLOps or Infrastructure role. While you will own the ML lifecycle, your day-
to-day focus will be heavily indexed on core machine learning methodology:
feature and target design, data exploration, training strong supervised
baselines, and rigorous evaluation on real-world tabular data.

Empower is a high-growth, innovation-driven company focused on delivering
impactful solutions that transform how healthcare organizations operate. Our
team combines deep industry expertise with a commitment to excellence,
enabling us to solve complex challenges at scale. We foster a collaborative,
results-oriented culture where high performers thrive, ideas are valued, and
continuous improvement is expected. We move quickly, think strategically, and
hold a high bar for performance, while supporting each other in achieving
ambitious goals.

The emphasis for this role is new modeling: experimentation, feature and
target design, training strong baselines and improved models, and rigorous
evaluation. You will partner with backend engineers for integration, but the
primary expectation is depth in ML methodology and healthcare-relevant
signals, not primarily platform or MLOps ownership.
We are focused on applying machine learning to solve meaningful, real-world
problems in a clinical healthcare industry. Our team partners closely with
product, engineering, and domain experts to translate data into actionable
insights and automated solutions. Machine Learning Engineers own the full
lifecycle - from data exploration and model development to deployment,
monitoring, and iteration in production.

Healthcare domain: DRG, ICD-10, appeals, utilization review, or EHR-derived
features. * _Preference will be given to candidates that demonstrate domain
knowledge and experience working on clincial workflow related models.
*_

Must Have

Applied Modeling Focus: Deep interest in the data math and methodology over
infrastructure. You are excited to roll up your sleeves to solve messy
clinical data problems through feature engineering and model accuracy,
partnering with our backend team who assists with the deployment tooling.

Core ML: Strong Python and hands-on experience training, evaluating, and
shipping supervised models (classification and related tasks). Comfortable
with scikit-learn-style pipelines, feature preparation, cross-validation, and
model diagnostics (calibration, drift concepts, error analysis).

Relevant experience: A track record of applied ML in industry or equivalent
(e.g. multiple shipped or production-adjacent projects). A PhD is not
required; demonstrated impact and judgment matter more than degree level.

Data: Experience with tabular data at scale. Ability to write or collaborate
on SQL and to reason about features from a data warehouse (we use Snowflake
heavily, including Snowpark and ML-related integrations). Healthcare or
regulated-domain experience is a strong plus.

Production awareness: Enough familiarity to collaborate on inference APIs,
batch jobs, and failure modes. You do not need to own the entire MLOps stack;
clarity on what "good enough to hand off" looks like is important.

Software engineering: Can read and contribute to a shared codebase (FastAPI
services, configuration, tests). Comfortable with Git, code review, and
documenting training assumptions and reproducibility.

Compliance awareness: Understanding that PHI and clinical workflows impose
constraints on logging, data retention, and access. Willingness to follow
security and privacy guidance from the broader team.

Nice to Have

Cloud ML: Azure Machine Learning or equivalent (SageMaker, Vertex).
Snowflake: Snowpark Python, warehouse-side processing, or SQL tuning for
feature extraction.
Orchestration: Temporal or similar for durable batch workflows.
LLMs and NLP in production: Ensemble or multi-model setups, clinical NLP, or
entity extraction patterns.

MLOps: Experiment tracking, automated retraining, or evaluation in CI
(supporting, not the main focus of the role).