Lead Applied Scientist, NBA – Reinforcement Learning

Humana

2d•$129,300 - $177,800•Hybrid

About The Position

Humana is seeking a Lead Applied Scientist to design, develop, and scale intelligent decision systems for its Next Best Action (NBA) platform. The platform uses machine learning, reinforcement learning, and decision intelligence to provide personalized recommendations intended to improve healthcare experiences and outcomes for millions of members. This is a hands-on technical leadership role translating complex business problems into scalable, production-ready decision systems. The scientist owns models and decision policies end to end, from prototyping and experimentation through offline evaluation, simulation, testing, production deployment, monitoring, and continuous improvement. The role evaluates system behavior under real-world constraints and uncertainty and improves performance using production outcomes. The position is remote or hybrid, preferably near Boston, Massachusetts. Occasional travel to Humana offices for training or meetings may be required. Typical hours are Monday through Friday, eight hours per day, with some flexibility depending on business needs. The role is not eligible for work visa sponsorship. Home or hybrid employees must have internet service with at least 25 Mbps download and 10 Mbps upload speeds and a dedicated workspace that protects member PHI/HIPAA information. Scheduled hours are 40 per week. The estimated starting base pay is $129,300–$177,800 per year, with individual pay dependent on factors including location, skills, experience, education, and certifications. The position is eligible for a bonus incentive plan based on company and/or individual performance. The application deadline is November 3, 2026.

Requirements

Bachelor’s degree in Computer Science, Engineering, Mathematics, Statistics, Operations Research, or a related quantitative field.
At least 6 years of professional experience in machine learning, applied science, decision science, or related engineering disciplines.
Strong experience designing, developing, testing, deploying, and monitoring machine learning solutions in production environments.
Solid understanding of reinforcement learning, sequential decision-making, or other advanced decision-making techniques.
Strong Python programming skills and experience with machine learning frameworks such as PyTorch or TensorFlow.
Experience with distributed data processing and machine learning platforms such as Databricks, PySpark, Ray, or equivalent technologies.
Strong software engineering fundamentals, including software architecture, code quality, automated testing, and scalable system design.
Ability to translate complex business problems into analytical models and production solutions.
Experience owning models across the full lifecycle, including prototyping, offline evaluation, simulation, testing, deployment, and production monitoring.

Nice To Haves

Experience with reinforcement learning algorithms, contextual bandits, Markov Decision Processes, or offline policy evaluation.
Experience with operations research, simulation, or decision analysis.
Familiarity with MLflow, Ray RLlib, Databricks, and distributed model training.
Experience developing recommendation systems, personalization platforms, or Next Best Action solutions.
Experience with event-driven architectures, Kafka, Redis, or rule-based decision engines.
Familiarity with decision mapping, multi-objective decision-making, and decision-making under uncertainty.
Experience monitoring deployed models and policies for changes in performance, data distributions, and decision behavior.
Experience within healthcare, insurance, or other regulated industries.

Responsibilities

Design, develop, and improve machine learning and reinforcement learning models for personalized recommendations, sequential decision-making, and policy learning.
Apply decision science, operations research, and simulation techniques to formulate complex decision problems, evaluate alternatives, and improve outcomes under business and operational constraints.
Develop and improve production machine learning pipelines using Python, PyTorch, Databricks, PySpark, Ray RLlib, and MLflow for large-scale data processing, model training, and inference.
Design experiments, simulations, offline evaluations, and performance metrics to assess model effectiveness, policy behavior, decision quality, and business impact before and after deployment.
Own models and decision policies throughout their lifecycle, from prototyping and testing through production deployment, monitoring, and continuous improvement.
Partner with software engineering and platform teams to ensure production systems are scalable, reliable, and observable.
Incorporate clinical eligibility, member preferences, business rules, operational capacity, and other constraints into intelligent decision systems.
Provide technical direction on model architecture, algorithm selection, evaluation methods, engineering best practices, and system design.
Collaborate with data scientists, engineers, and business stakeholders.