Job Description

Job Summary
The Comprehensive Mobile Precision Approach for Scalable Solutions in Mental Health Treatment (COMPASS) project is looking for an Application System Analyst/Programmer Intermediate for our new NIMH project. COMPASS is focused on the effects of machine learning, wearable and mobile technology to both reduce mental health symptoms and predict response to clinic-based treatments. The ideal candidate should be passionate about building data and software infrastructure to power precision mental health research that integrates mobile sensing, electronic health records, and reinforcement learning. We are looking for someone with a demonstrated background in programming who is detail-oriented and will work to conceptualize, develop and implement complex program designs to assist with the reinforcement learning AI agent to coalesce complex data from multiple sources (wearable devices, smartphones, survey answers).

Responsibilities*
Responsibilities include, but are not limited to:

Reinforcement Learning Algorithm Development and Implementation (35%)

Lead the design, implementation, and operation of reinforcement learning algorithms.
Contribute to study design discussions, translate research questions into technical requirements, and communicate options, tradeoffs, and timelines to stakeholders.
Uphold data governance and research compliance, including HIPAA and data use agreements, and support secure data sharing with partner institutions.
Track emerging methods and tools in the reinforcement learning field and evaluate and deploy those that improve performance, cost, or scientific value.
Use Large Language Models to Generate Intervention Messages (25%)

Lead the development of a sustainable infrastructure for generating intervention messages to deploy to study participants on a large scale across multiple projects.
Design an interface for collecting and integrating study team feedback in the message creation process.
Code and documentation development (15%)

Develop and improve processes for extracting, transforming and loading various complex data sources (reference data sets, EHR, etc.) into study data pipeline.
Build, maintain, and update databases/scripts/programs aimed at improving data cleaning processes by writing automatic scripts and templates (using Python, R, JavaScript, bash shell, or equivalent).
Develop data documentation (codebooks, technical appendices, etc.) that includes key information for processing and analysis following standard guidelines.
Identify opportunities to improve the efficiency and quality of data processing and analysis and collaborate with team members to implement solutions.
Implement monitoring solutions to ensure data integrity and pipelines are operational.
Investigate new technologies and share with the team.

Data Team work (15%)

Participate as a team member in weekly discussions on analysis and improvement of data collection, quality of data analyses, programming and documentation.
Communicate proactively with project managers and PIs about data progress, timelines, and issues.
Identify and resolve time-sensitive data issues.
Document reinforcement learning process.
Manuscript Development (10%)

Develop data visualizations to effectively communicate data in reports, flowcharts, and dashboards.
Support the writing team by preparing figures, methods, and reproducible analyses; coauthor manuscripts and lead selected method or application papers, including drafting, submission, and revisions.
Support research design, determining and interpreting research results, and independently writing up results of analyses.
Required Qualifications*
Master's degree in Biostatistics
Demonstrated background in reinforcement learning with proven experience in building adaptive, sequential decision systems for health research.
Reinforcement learning-based contextual bandits in applied settings.
Integrating wearable, smartphone, survey, and longitudinal behavioral data.
Designing algorithms and workflows to implement existing theoretical literature within an online, clinical setting.
GPU-based statistical computations.
Local LLM data pipelines.
Proficiency using relevant tools, including Python, R, Git, and version-controlled scientific workflows.
Background in database design and server-based computing.
Clear, professional communication skills, with the ability to translate between technical and scientific audiences; demonstrated record of collaboration and independent problem solving.
Willingness to learn and quickly adjust to new tools and systems.
Ability to manage multiple projects and assignments with a high level of autonomy and accountability for results.
Experience in mentoring and supervising the work of a team, peers, and staff.
Capable of converting ambiguous problem statements into concrete project requirements.
Ability to understand and explain technical concepts to non-technical stakeholders.
Desired Qualifications*
Experience with cloud platforms (for example, GCP or AWS).
Familiar with Michigan Medicine data requirements and procedures, including EPIC Electronic Health Records.
Experience working with big data.
Experience working with PHI.